From c16e2133d8c1b9505752e2c8f4e2b4d0e1248909 Mon Sep 17 00:00:00 2001 From: Adam Harvey Date: Wed, 27 Feb 2019 14:58:51 +0100 Subject: update .md --- site/content/pages/about/assets/adam-harvey.jpg | Bin 0 -> 18525 bytes site/content/pages/about/assets/jules-laplace.jpg | Bin 0 -> 15254 bytes site/content/pages/about/index.md | 26 ++++++++++++------ site/content/pages/about/press.md | 1 + site/content/pages/datasets/caltech_10k/index.md | 29 +++++++++++++++++++++ site/content/pages/datasets/lfw/index.md | 4 ++- site/content/pages/info/index.md | 2 +- .../pages/research/00_introduction/index.md | 20 +++++++------- 8 files changed, 62 insertions(+), 20 deletions(-) create mode 100644 site/content/pages/about/assets/adam-harvey.jpg create mode 100644 site/content/pages/about/assets/jules-laplace.jpg create mode 100644 site/content/pages/datasets/caltech_10k/index.md (limited to 'site') diff --git a/site/content/pages/about/assets/adam-harvey.jpg b/site/content/pages/about/assets/adam-harvey.jpg new file mode 100644 index 00000000..e0ab893a Binary files /dev/null and b/site/content/pages/about/assets/adam-harvey.jpg differ diff --git a/site/content/pages/about/assets/jules-laplace.jpg b/site/content/pages/about/assets/jules-laplace.jpg new file mode 100644 index 00000000..310b2783 Binary files /dev/null and b/site/content/pages/about/assets/jules-laplace.jpg differ diff --git a/site/content/pages/about/index.md b/site/content/pages/about/index.md index e2025bf2..f9c6f83a 100644 --- a/site/content/pages/about/index.md +++ b/site/content/pages/about/index.md @@ -1,8 +1,8 @@ ------------ status: published -title: MegaPixels Credits -desc: MegaPixels Project Team Credits +title: About MegaPixels +desc: About MegaPixels slug: credits published: 2018-12-04 updated: 2018-12-04 @@ -10,10 +10,20 @@ authors: Adam Harvey ------------ -# Credits +# About MegaPixels -- MegaPixels by Adam Harvey -- Made with support from Mozilla -- Site developed by Jules Laplace -- Design and graphics: Adam Harvey -- Research assistants: Berit Gilma \ No newline at end of file +MegaPixels aims to answers to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla. + +MegaPixels aims to answers to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla. + ++ Years: 2002-2004 ++ Datasets Analyzed: 325 ++ Author: Adam Harvey ++ Development: Jules LaPlace ++ Research Assistance: Berit Gilma + +![Adam Harvey](assets/adam-harvey.jpg) **Adam Harvey** is an American artist and researcher based in Berlin. His previous projects (CV Dazzle, Stealth Wear, and SkyLift) explore the potential for countersurveillance as artwork. He is the founder of VFRAME (visual forensics software for human rights groups), the recipient of 2 PrototypeFund awards, and is currently a researcher in residence at Karlsruhe HfG studying artifical intelligence and datasets. + +![Adam Harvey](assets/jules-laplace.jpg) **Jules LaPlace** is an American technologist and artist also based in Berlin. He was previously the CTO for a NYC digital agency and currently works at VFRAME, developing computer vision for human rights groups, and as a freelance technologists for artists. + +**Mozilla** is a free software community founded in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions. The community is supported institutionally by the not-for-profit Mozilla Foundation and its tax-paying subsidiary, the Mozilla Corporation. \ No newline at end of file diff --git a/site/content/pages/about/press.md b/site/content/pages/about/press.md index 56b4990f..2e3fa9a7 100644 --- a/site/content/pages/about/press.md +++ b/site/content/pages/about/press.md @@ -18,3 +18,4 @@ authors: Adam Harvey - Aug 22, 2018: "Transgender YouTubers had their videos grabbed to train facial recognition software" by James Vincent - Aug 22, 2018: "Transgender YouTubers had their videos grabbed to train facial recognition software" by James Vincent - Aug 22, 2018: "Transgender YouTubers had their videos grabbed to train facial recognition software" by James Vincent +lfw \ No newline at end of file diff --git a/site/content/pages/datasets/caltech_10k/index.md b/site/content/pages/datasets/caltech_10k/index.md new file mode 100644 index 00000000..8f49f2d1 --- /dev/null +++ b/site/content/pages/datasets/caltech_10k/index.md @@ -0,0 +1,29 @@ +------------ + +status: published +title: Caltech 10K Faces Dataset +desc: Caltech 10K Faces Dataset +slug: caltech_10k +published: 2019-2-23 +updated: 2019-2-23 +authors: Adam Harvey + +------------ + +# Caltech 10K Faces Dataset + ++ Years: TBD ++ Images: TBD ++ Identities: TBD ++ Origin: Google Search ++ Funding: TBD + +------- + +Ignore text below these lines + +------- + +Research + +The dataset contains images of people collected from the web by typing common given names into Google Image Search. The coordinates of the eyes, the nose and the center of the mouth for each frontal face are provided in a ground truth file. This information can be used to align and crop the human faces or as a ground truth for a face detection algorithm. The dataset has 10,524 human faces of various resolutions and in different settings, e.g. portrait images, groups of people, etc. Profile faces or very low resolution faces are not labeled. \ No newline at end of file diff --git a/site/content/pages/datasets/lfw/index.md b/site/content/pages/datasets/lfw/index.md index 1f847a2a..8b37f035 100644 --- a/site/content/pages/datasets/lfw/index.md +++ b/site/content/pages/datasets/lfw/index.md @@ -2,10 +2,12 @@ status: published title: Labeled Faces in The Wild -desc: LFW: Labeled Faces in The Wild +desc: Labeled Faces in The Wild (LFW) is a database of face photographs designed for studying the problem of unconstrained face recognition +subdesc: It includes 13,456 images of 4,432 people’s images copied from the Internet during 2002-2004. slug: lfw published: 2019-2-23 updated: 2019-2-23 +color: #00FF00 authors: Adam Harvey ------------ diff --git a/site/content/pages/info/index.md b/site/content/pages/info/index.md index 4a65e71a..9cbb219e 100644 --- a/site/content/pages/info/index.md +++ b/site/content/pages/info/index.md @@ -11,7 +11,7 @@ sync: false ------------ -## What do facial recognition algorithms see? +## ``` face_analysis diff --git a/site/content/pages/research/00_introduction/index.md b/site/content/pages/research/00_introduction/index.md index d3ef506b..1b784768 100644 --- a/site/content/pages/research/00_introduction/index.md +++ b/site/content/pages/research/00_introduction/index.md @@ -15,6 +15,16 @@ authors: Megapixels + Posted: Dec. 15 + Author: Adam Harvey + +----- + +Ignore content below these lines + +----- + +Ever since the first computational facial recognition research project by the CIA in the early 1960s, data has always played a vital role in the development of our biometric future. Without facial recognition datasets there would be no facial recognition. Datasets are an indispensable part of any artificial intelligence system because, as Geoffrey Hinton points out, "we no longer program computers with code, we program them with data". + + It was the early 2000s. Face recognition was new and no one seemed sure exactly how well it was going to perform in practice. In theory, face recognition was poised to be a game changer, a force multiplier, a strategic military advantage, a way to make cities safer and to secure borders. This was the future John Ashcroft demanded with the Total Information Awareness act of the 2003 and that spooks had dreamed of for decades. It was a future that academics at Carnegie Mellon Universtiy and Colorado State University would help build. It was also a future that celebrities would play a significant role in building. And to the surprise of ordinary Internet users like myself and perhaps you, it was a future that millions of Internet users would unwittingly play role in creating. Now the future has arrived and it doesn't make sense. Facial recognition works yet it doesn't actually work. Facial recognition is cheap and accessible but also expensive and out of control. Facial recognition research has achieved headline grabbing superhuman accuracies over 99.9% yet facial recognition is also dangerously inaccurate. During a trial installation at Sudkreuz station in Berlin in 2018, 20% of the matches were wrong, a number so low that it should not have any connection to law enforcement or justice. And in London, the Metropolitan police had been using facial recognition software that mistakenly identified an alarming 98% of people as criminals [^met_police], which perhaps is a crime itself. @@ -33,16 +43,6 @@ As McLuhan wrote, "You can't have a static, fixed position in the electric age". Like many projects, MegaPixels had spent years meandering between formats, unfeasible budgets, and was generally too niche of a subject. The basic idea for this project, as proposed to the original [Glass Room](https://tacticaltech.org/projects/the-glass-room-nyc/) installation in 2016 in NYC, was to build an interactive mirror that showed people if they had been included in the [LFW](/datasets/lfw) facial recognition dataset. The idea was based on my reaction to all the datasets I'd come across during research for the CV Dazzle project. I'd noticed strange datasets created for training and testing face detection algorithms. Most were created in labratory settings and their interpretation of face data was very strict. -About the name - -About the funding - -About me - -About the team - -Conclusion - ### for other post -- cgit v1.2.3-70-g09d2 From c8e7a10be948c2405d46d8c3caf4a8c6675eee29 Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Wed, 27 Feb 2019 19:35:54 +0100 Subject: rebuild --- client/map/index.js | 1 + megapixels/app/settings/app_cfg.py | 2 +- megapixels/app/site/builder.py | 2 +- megapixels/app/site/parser.py | 3 +- site/assets/css/css.css | 7 +- site/public/about/index.html | 14 +-- site/public/about/press/index.html | 3 +- site/public/datasets/lfw/index.html | 118 +++++++++++-------- site/public/datasets/vgg_face2/index.html | 33 +----- site/public/datasets_v0/index.html | 53 +++++++++ site/public/datasets_v0/lfw/index.html | 131 +++++++++++++++++++++ .../datasets_v0/lfw/right-to-removal/index.html | 62 ++++++++++ site/public/datasets_v0/lfw/tables/index.html | 52 ++++++++ site/public/datasets_v0/vgg_face2/index.html | 80 +++++++++++++ site/public/index.html | 50 +++++++- site/public/info/index.html | 2 +- site/public/research/00_introduction/index.html | 9 +- 17 files changed, 519 insertions(+), 103 deletions(-) create mode 100644 site/public/datasets_v0/index.html create mode 100644 site/public/datasets_v0/lfw/index.html create mode 100644 site/public/datasets_v0/lfw/right-to-removal/index.html create mode 100644 site/public/datasets_v0/lfw/tables/index.html create mode 100644 site/public/datasets_v0/vgg_face2/index.html (limited to 'site') diff --git a/client/map/index.js b/client/map/index.js index 2a6686be..d38855bf 100644 --- a/client/map/index.js +++ b/client/map/index.js @@ -78,6 +78,7 @@ export default function append(el, payload) { source = [address.lat, address.lng].map(n => parseFloat(n)) } + // ....i dont think the sort order does anything?? citations.sort((a,b) => sortOrder.indexOf(a) - sortOrder.indexOf(b)) .forEach(citation => { const address = citation.addresses[0] diff --git a/megapixels/app/settings/app_cfg.py b/megapixels/app/settings/app_cfg.py index 0b1fb69d..40625958 100644 --- a/megapixels/app/settings/app_cfg.py +++ b/megapixels/app/settings/app_cfg.py @@ -163,7 +163,7 @@ S3_HTTP_METADATA_URL = join(S3_HTTP_URL, 'metadata') S3_SITE_PATH = "v1/site" S3_DATASETS_PATH = "v1" # datasets is already in the filename DIR_SITE_PUBLIC = "../site/public" -DIR_SITE_CONTENT = "../site/content" +DIR_SITE_CONTENT = "../site/content/pages" DIR_SITE_TEMPLATES = "../site/templates" DIR_SITE_USER_CONTENT = "../site/public/user_content" diff --git a/megapixels/app/site/builder.py b/megapixels/app/site/builder.py index fac49c24..188fbc25 100644 --- a/megapixels/app/site/builder.py +++ b/megapixels/app/site/builder.py @@ -73,7 +73,7 @@ def build_index(key, research_posts, datasets): """ build the index of research (blog) posts """ - metadata, sections = parser.read_metadata('../site/content/{}/index.md'.format(key)) + metadata, sections = parser.read_metadata(os.path.join(cfg.DIR_SITE_CONTENT, key, 'index.md')) template = env.get_template("page.html") s3_path = s3.make_s3_path(cfg.S3_SITE_PATH, metadata['path']) content = parser.parse_markdown(sections, s3_path, skip_h1=False) diff --git a/megapixels/app/site/parser.py b/megapixels/app/site/parser.py index f739315a..d6705214 100644 --- a/megapixels/app/site/parser.py +++ b/megapixels/app/site/parser.py @@ -127,6 +127,7 @@ def parse_research_index(research_posts): """ content = "
" for post in research_posts: + print(post) s3_path = s3.make_s3_path(cfg.S3_SITE_PATH, post['path']) if 'image' in post: post_image = s3_path + post['image'] @@ -240,7 +241,7 @@ def read_post_index(basedir): Generate an index of posts """ posts = [] - for fn in sorted(glob.glob('../site/content/{}/*/index.md'.format(basedir))): + for fn in sorted(glob.glob(os.path.join(cfg.DIR_SITE_CONTENT, basedir, '*/index.md'))): metadata, valid_sections = read_metadata(fn) if metadata is None or metadata['status'] == 'private' or metadata['status'] == 'draft': continue diff --git a/site/assets/css/css.css b/site/assets/css/css.css index 7544fd9d..858d98eb 100644 --- a/site/assets/css/css.css +++ b/site/assets/css/css.css @@ -185,7 +185,7 @@ th, .gray { line-height: 1.5; } section { - width: 640px; + width: 960px; margin: 0 auto; } .home section { @@ -251,6 +251,7 @@ ul { ul li { margin-bottom: 8px; } + /* misc formatting */ code { @@ -267,7 +268,7 @@ pre { pre code { display: block; max-height: 400px; - max-width: 640px; + max-width: 960px; overflow: scroll; padding: 4px 10px; } @@ -416,7 +417,7 @@ section.fullwidth .image { font-size: 26px; } .intro { - max-width: 640px; + max-width: 960px; padding: 75px 0 75px 10px; z-index: 1; } diff --git a/site/public/about/index.html b/site/public/about/index.html index fecc6c7b..4a5ca926 100644 --- a/site/public/about/index.html +++ b/site/public/about/index.html @@ -4,7 +4,7 @@ MegaPixels - + @@ -27,14 +27,10 @@
-

Credits

-
    -
  • MegaPixels by Adam Harvey
    • -
  • Made with support from Mozilla
    • -
  • Site developed by Jules Laplace
    • -
  • Design and graphics: Adam Harvey
    • -
  • Research assistants: Berit Gilma
    • -
+

About MegaPixels

+

MegaPixels aims to answers to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla.

+

MegaPixels aims to answers to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla.

+
Years
2002-2004
Datasets Analyzed
325
Author
Adam Harvey
Development
Jules LaPlace
Research Assistance
Berit Gilma
Adam Harvey
Adam Harvey
Adam Harvey
Adam Harvey

Mozilla is a free software community founded in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions. The community is supported institutionally by the not-for-profit Mozilla Foundation and its tax-paying subsidiary, the Mozilla Corporation.

diff --git a/site/public/about/press/index.html b/site/public/about/press/index.html index b9dd97c2..a1d9d4f5 100644 --- a/site/public/about/press/index.html +++ b/site/public/about/press/index.html @@ -31,7 +31,8 @@
alt text
alt text
diff --git a/site/public/datasets/lfw/index.html b/site/public/datasets/lfw/index.html index a6226720..f83d8a66 100644 --- a/site/public/datasets/lfw/index.html +++ b/site/public/datasets/lfw/index.html @@ -4,7 +4,7 @@ MegaPixels - + @@ -27,54 +27,60 @@
-

Labeled Faces in the Wild

-
Created
2007
Images
13,233
People
5,749
Created From
Yahoo News images
Search available
Searchable
Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.
Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.

Intro

-

Labeled Faces in The Wild (LFW) is among the most widely used facial recognition training datasets in the world and is the first of its kind to be created entirely from images posted online. The LFW dataset includes 13,233 images of 5,749 people that were collected between 2002-2004. Use the tools below to check if you were included in this dataset or scroll down to read the analysis.

-

Three paragraphs describing the LFW dataset in a format that can be easily replicated for the other datasets. Nothing too custom. An analysis of the initial research papers with context relative to all the other dataset papers.

-
From George W. Bush to Jamie Lee Curtis: all 5,749 people in the LFW Dataset sorted from most to least images collected.
From George W. Bush to Jamie Lee Curtis: all 5,749 people in the LFW Dataset sorted from most to least images collected.

LFW by the Numbers

+

LFW

+
Years
2002-2004
Images
13,233
Identities
5,749
Origin
Yahoo News Images
Funding
(Possibly, partially CIA*)
Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.
Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.

Labeled Faces in The Wild (LFW) is "a database of face photographs designed for studying the problem of unconstrained face recognition[^lfw_www]. It is used to evaluate and improve the performance of facial recognition algorithms in academic, commercial, and government research. According to BiometricUpdate.com[^lfw_pingan], LFW is "the most widely used evaluation set in the field of facial recognition, LFW attracts a few dozen teams from around the globe including Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong."

+

The LFW dataset includes 13,233 images of 5,749 people that were collected between 2002-2004. LFW is a subset of Names of Faces and is part of the first facial recognition training dataset created entirely from images appearing on the Internet. The people appearing in LFW are...

+

The Names and Faces dataset was the first face recognition dataset created entire from online photos. However, Names and Faces and LFW are not the first face recognition dataset created entirely "in the wild". That title belongs to the UCD dataset. Images obtained "in the wild" means using an image without explicit consent or awareness from the subject or photographer.

+

Analysis

    -
  • Was first published in 2007
    • -
  • Developed out of a prior dataset from Berkely called "Faces in the Wild" or "Names and Faces" [^lfw_original_paper]
    • -
  • Includes 13,233 images and 5,749 different people [^lfw_website]
    • -
  • There are about 3 men for every 1 woman (4,277 men and 1,472 women)[^lfw_website]
    • -
  • The person with the most images is George W. Bush with 530
    • -
  • Most people (70%) in the dataset have only 1 image
    • -
  • Thre are 1,680 people in the dataset with 2 or more images [^lfw_website]
    • -
  • Two out of 4 of the original authors received funding from the Office of Director of National Intelligence and IARPA for their 2016 LFW survey follow up report
    • -
  • The LFW dataset includes over 500 actors, 30 models, 10 presidents, 24 football players, 124 basketball players, 11 kings, and 2 queens
    • -
  • In all the LFW publications provided by the authors the words "ethics", "consent", and "privacy" appear 0 times [^lfw_original_paper], [^lfw_survey], [^lfw_tech_report] , [^lfw_website]
    • +
  • There are about 3 men for every 1 woman (4,277 men and 1,472 women) in the LFW dataset[^lfw_www]
    • +
  • The person with the most images is George W. Bush with 530
    • +
  • There are about 3 George W. Bush's for every 1 Tony Blair
    • +
  • 70% of people in the dataset have only 1 image and 29% have 2 or more images
    • +
  • The LFW dataset includes over 500 actors, 30 models, 10 presidents, 124 basketball players, 24 football players, 11 kings, 7 queens, and 1 Moby
    • +
  • In all 3 of the LFW publications [^lfw_original_paper], [^lfw_survey], [^lfw_tech_report] the words "ethics", "consent", and "privacy" appear 0 times
  • The word "future" appears 71 times
-

Facts

+

Synthetic Faces

+

To visualize the types of photos in the dataset without explicitly publishing individual's identities a generative adversarial network (GAN) was trained on the entire dataset. The images in this video show a neural network learning the visual latent space and then interpolating between archetypical identities within the LFW dataset.

+

Biometric Trade Routes

+

To understand how this dataset has been used, its citations have been geocoded to show an approximate geographic digital trade route of the biometric data. Lines indicate an organization (education, commercial, or governmental) that has cited the LFW dataset in their research. Data is compiled from SemanticScholar.

+

[add map here]

+

Citations

+

Browse or download the geocoded citation data collected for the LFW dataset.

+

[add citations table here]

+

Additional Information

+

(tweet-sized snippets go here)

    -
  • Was created for the purpose of improving "unconstrained face recognition" [^lfw_original_paper]
    • -
  • All images in LFW were obtained "in the wild" meaning without any consent from the subject or from the photographer
    • -
  • The faces were detected using the Viola-Jones haarcascade face detector [^lfw_website] [^lfw_survey]
    • -
  • Is considered the "most popular benchmark for face recognition" [^lfw_baidu]
    • -
  • Is "the most widely used evaluation set in the field of facial recognition" [^lfw_pingan]
    • -

  • Is used by several of the largest tech companies in the world including "Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong." [^lfw_pingan]

    -
    • -

  • All images were copied from Yahoo News between 2002 - 2004 [^lfw_original_paper]

    +
  • The LFW dataset is considered the "most popular benchmark for face recognition" [^lfw_baidu]
    • +
  • The LFW dataset is "the most widely used evaluation set in the field of facial recognition" [^lfw_pingan]
    • +
  • All images in LFW dataset were obtained "in the wild" meaning without any consent from the subject or from the photographer
    • +
  • The faces in the LFW dataset were detected using the Viola-Jones haarcascade face detector [^lfw_website] [^lfw-survey]
    • +
  • The LFW dataset is used by several of the largest tech companies in the world including "Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong." [^lfw_pingan]
    • +
  • All images in the LFW dataset were copied from Yahoo News between 2002 - 2004 +<<<<<<< HEAD
    • +
  • In 2014, two of the four original authors of the LFW dataset received funding from IARPA and ODNI for their follow up paper Labeled Faces in the Wild: Updates and New Reporting Procedures via IARPA contract number 2014-14071600010
    • +

  • The dataset includes 2 images of George Tenet, the former Director of Central Intelligence (DCI) for the Central Intelligence Agency whose facial biometrics were eventually used to help train facial recognition software in China and Russia

    • -
  • SenseTime, who has relied on LFW for benchmarking their facial recognition performance, is the leading provider of surveillance to the Chinese Government
    • +
  • In 2014, 2/4 of the original authors of the LFW dataset received funding from IARPA and ODNI for their follow up paper "Labeled Faces in the Wild: Updates and New Reporting Procedures" via IARPA contract number 2014-14071600010
    • +
  • The LFW dataset was used Center for Intelligent Information Retrieval, the Central Intelligence Agency, the National Security Agency and National
+

TODO (need citations for the following)

+
    +
  • SenseTime, who has relied on LFW for benchmarking their facial recognition performance, is one the leading provider of surveillance to the Chinese Government [need citation for this fact. is it the most? or is that Tencent?]
    • +
  • Two out of 4 of the original authors received funding from the Office of Director of National Intelligence and IARPA for their 2016 LFW survey follow up report
    • +
+

> 13d7a450affe8ea4f368a97ea2014faa17702a4c

+
+
+
+
+
+
+
former President George W. Bush
former President George W. Bush
-
Colin Powell (236), Tony Blair (144), and Donald Rumsfeld (121)
Colin Powell (236), Tony Blair (144), and Donald Rumsfeld (121)

People and Companies using the LFW Dataset

-

This section describes who is using the dataset and for what purposes. It should include specific examples of people or companies with citations and screenshots. This section is followed up by the graph, the map, and then the supplementary material.

-

The LFW dataset is used by numerous companies for benchmarking algorithms and in some cases training. According to the benchmarking results page [^lfw_results] provided by the authors, over 2 dozen companies have contributed their benchmark results.

-

According to BiometricUpdate.com [^lfw_pingan], LFW is "the most widely used evaluation set in the field of facial recognition, LFW attracts a few dozen teams from around the globe including Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong."

-

According to researchers at the Baidu Research – Institute of Deep Learning "LFW has been the most popular evaluation benchmark for face recognition, and played a very important role in facilitating the face recognition society to improve algorithm. [^lfw_baidu]."

-

In addition to commercial use as an evaluation tool, alll of the faces in LFW dataset are prepackaged into a popular machine learning code framework called scikit-learn.

-
"PING AN Tech facial recognition receives high score in latest LFW test results"
"PING AN Tech facial recognition receives high score in latest LFW test results"
-
"Face Recognition Performance in LFW benchmark"
"Face Recognition Performance in LFW benchmark"
-
"The 1st place in face verification challenge, LFW"
"The 1st place in face verification challenge, LFW"

In benchmarking, companies use a dataset to evaluate their algorithms which are typically trained on other data. After training, researchers will use LFW as a benchmark to compare results with other algorithms.

-

For example, Baidu (est. net worth $13B) uses LFW to report results for their "Targeting Ultimate Accuracy: Face Recognition via Deep Embedding". According to the three Baidu researchers who produced the paper:

-

Citations

-

Overall, LFW has at least 116 citations from 11 countries.

-

Conclusion

-

The LFW face recognition training and evaluation dataset is a historically important face dataset as it was the first popular dataset to be created entirely from Internet images, paving the way for a global trend towards downloading anyone’s face from the Internet and adding it to a dataset. As will be evident with other datasets, LFW’s approach has now become the norm.

-

For all the 5,000 people in this datasets, their face is forever a part of facial recognition history. It would be impossible to remove anyone from the dataset because it is so ubiquitous. For their rest of the lives and forever after, these 5,000 people will continue to be used for training facial recognition surveillance.

-

Code

+
Colin Powell (236), Tony Blair (144), and Donald Rumsfeld (121)
Colin Powell (236), Tony Blair (144), and Donald Rumsfeld (121)
All 5,379 faces in the Labeled Faces in The Wild Dataset
All 5,379 faces in the Labeled Faces in The Wild Dataset

Code

+

The LFW dataset is so widely used that a popular code library called Sci-Kit Learn includes a function called fetch_lfw_people to download the faces in the LFW dataset.

#!/usr/bin/python
 
 import numpy as np
@@ -87,26 +93,38 @@ lfw_people = fetch_lfw_people(min_faces_per_person=1, resize=1, color=True, funn
 
 # introspect dataset
 n_samples, h, w, c = lfw_people.images.shape
-print('{:,} images at {}x{}'.format(n_samples, w, h))
+print(f'{n_samples:,} images at {w}x{h} pixels')
 cols, rows = (176, 76)
 n_ims = cols * rows
 
 # build montages
 im_scale = 0.5
-ims = lfw_people.images[:n_ims
-montages = imutils.build_montages(ims, (int(w*im_scale, int(h*im_scale)), (cols, rows))
+ims = lfw_people.images[:n_ims]
+montages = imutils.build_montages(ims, (int(w * im_scale,   int(h * im_scale)), (cols, rows))
 montage = montages[0]
 
 # save full montage image
 imageio.imwrite('lfw_montage_full.png', montage)
 
 # make a smaller version
-montage_960 = imutils.resize(montage, width=960)
-imageio.imwrite('lfw_montage_960.jpg', montage_960)
+montage = imutils.resize(montage, width=960)
+imageio.imwrite('lfw_montage_960.jpg', montage)
-

Disclaimer

-

MegaPixels is an educational art project designed to encourage discourse about facial recognition datasets. Any ethical or legal issues should be directed to the researcher's parent organizations. Except where necessary for contact or clarity, the names of researchers have been subsituted by their parent organization. In no way does this project aim to villify researchers who produced the datasets.

-

Read more about MegaPixels Code of Conduct

+

Supplementary Material

+

Text and graphics ©Adam Harvey / megapixels.cc

+

Ignore text below these lines

+

Research

+
    +
  • "In our experiments, we used 10000 images and associated captions from the Faces in the wilddata set [3]."
    • +
  • "This work was supported in part by the Center for Intelligent Information Retrieval, the Central Intelligence Agency, the National Security Agency and National Science Foundation under CAREER award IIS-0546666 and grant IIS-0326249."
    • +
  • From: "People-LDA: Anchoring Topics to People using Face Recognition" https://www.semanticscholar.org/paper/People-LDA%3A-Anchoring-Topics-to-People-using-Face-Jain-Learned-Miller/10f17534dba06af1ddab96c4188a9c98a020a459 and https://ieeexplore.ieee.org/document/4409055
    • +
  • This paper was presented at IEEE 11th ICCV conference Oct 14-21 and the main LFW paper "Labeled Faces in the Wild: A Database for Studying Face Recognition in Unconstrained Environments" was also published that same year
    • +

  • 10f17534dba06af1ddab96c4188a9c98a020a459

    +
    • +

  • This research is based upon work supported in part by the Office of the Director of National Intelligence (ODNI), Intelligence Advanced Research Projects Activity (IARPA), via contract number 2014-14071600010.

    +
    • +
  • From "Labeled Faces in the Wild: Updates and New Reporting Procedures"
    • +
    diff --git a/site/public/datasets/vgg_face2/index.html b/site/public/datasets/vgg_face2/index.html index b7ba5a4c..08b02cc7 100644 --- a/site/public/datasets/vgg_face2/index.html +++ b/site/public/datasets/vgg_face2/index.html @@ -4,7 +4,7 @@ MegaPixels - + @@ -27,35 +27,10 @@
    -

    VGG Faces2

    -
    Created
    2018
    Images
    3.3M
    People
    9,000
    Created From
    Scraping search engines
    Search available
    [Searchable](#)

    VGG Face2 is the updated version of the VGG Face dataset and now includes over 3.3M face images from over 9K people. The identities were selected by taking the top 500K identities in Google's Knowledge Graph of celebrities and then selecting only the names that yielded enough training images. The dataset was created in the UK but funded by Office of Director of National Intelligence in the United States.

    -

    VGG Face2 by the Numbers

    +

    VGG Face 2

    +
    Years
    TBD
    Images
    TBD
    Identities
    TBD
    Origin
    TBD
    Funding
    IARPA
    ...
    ...

    Analysis

      -
    • 1,331 actresses, 139 presidents
      • -
    • 3 husbands and 16 wives
      • -
    • 2 snooker player
      • -
    • 1 guru
      • -
    • 1 pornographic actress
      • -
    • 3 computer programmer
      • -
    -

    Names and descriptions

    -
      -
    • The original VGGF2 name list has been updated with the results returned from Google Knowledge
      • -
    • Names with a similarity score greater than 0.75 where automatically updated. Scores computed using import difflib; seq = difflib.SequenceMatcher(a=a.lower(), b=b.lower()); score = seq.ratio()
      • -
    • The 97 names with a score of 0.75 or lower were manually reviewed and includes name changes validating using Wikipedia.org results for names such as "Bruce Jenner" to "Caitlyn Jenner", spousal last-name changes, and discretionary changes to improve search results such as combining nicknames with full name when appropriate, for example changing "Aleksandar Petrović" to "Aleksandar 'Aco' Petrović" and minor changes such as "Mohammad Ali" to "Muhammad Ali"
      • -
    • The 'Description' text was automatically added when the Knowledge Graph score was greater than 250
      • -
    -

    TODO

    -
      -
    • create name list, and populate with Knowledge graph information like LFW
      • -
    • make list of interesting number stats, by the numbers
      • -
    • make list of interesting important facts
      • -
    • write intro abstract
      • -
    • write analysis of usage
      • -
    • find examples, citations, and screenshots of useage
      • -
    • find list of companies using it for table
      • -
    • create montages of the dataset, like LFW
      • -
    • create right to removal information
      • +
    • The VGG Face 2 dataset includes approximately 1,331 actresses, 139 presidents, 16 wives, 3 husbands, 2 snooker player, and 1 guru
    diff --git a/site/public/datasets_v0/index.html b/site/public/datasets_v0/index.html new file mode 100644 index 00000000..71147a64 --- /dev/null +++ b/site/public/datasets_v0/index.html @@ -0,0 +1,53 @@ + + + + MegaPixels + + + + + + + + + + + + +
    + + +
    MegaPixels
    +     +
    + Datasets + Research + About +
    +
    +
    + +

    Facial Recognition Datasets

    +

    Regular Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.

    +

    Summary

    +
    Found
    275 datasets
    Created between
    1993-2018
    Smallest dataset
    20 images
    Largest dataset
    10,000,000 images
    Highest resolution faces
    450x500 (Unconstrained College Students)
    Lowest resolution faces
    16x20 pixels (QMUL SurvFace)
    + +
    + + + + + \ No newline at end of file diff --git a/site/public/datasets_v0/lfw/index.html b/site/public/datasets_v0/lfw/index.html new file mode 100644 index 00000000..b4ee82a3 --- /dev/null +++ b/site/public/datasets_v0/lfw/index.html @@ -0,0 +1,131 @@ + + + + MegaPixels + + + + + + + + + + + + +
    + + +
    MegaPixels
    +     +
    + Datasets + Research + About +
    +
    +
    + +

    Labeled Faces in the Wild

    +
    Created
    2007
    Images
    13,233
    People
    5,749
    Created From
    Yahoo News images
    Search available
    Searchable
    Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.
    Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.

    Intro

    +

    Labeled Faces in The Wild (LFW) is among the most widely used facial recognition training datasets in the world and is the first of its kind to be created entirely from images posted online. The LFW dataset includes 13,233 images of 5,749 people that were collected between 2002-2004. Use the tools below to check if you were included in this dataset or scroll down to read the analysis.

    +

    Three paragraphs describing the LFW dataset in a format that can be easily replicated for the other datasets. Nothing too custom. An analysis of the initial research papers with context relative to all the other dataset papers.

    +
    From George W. Bush to Jamie Lee Curtis: all 5,749 people in the LFW Dataset sorted from most to least images collected.
    From George W. Bush to Jamie Lee Curtis: all 5,749 people in the LFW Dataset sorted from most to least images collected.

    LFW by the Numbers

    +
      +
    • Was first published in 2007
      • +
    • Developed out of a prior dataset from Berkely called "Faces in the Wild" or "Names and Faces" [^lfw_original_paper]
      • +
    • Includes 13,233 images and 5,749 different people [^lfw_website]
      • +
    • There are about 3 men for every 1 woman (4,277 men and 1,472 women)[^lfw_website]
      • +
    • The person with the most images is George W. Bush with 530
      • +
    • Most people (70%) in the dataset have only 1 image
      • +
    • Thre are 1,680 people in the dataset with 2 or more images [^lfw_website]
      • +
    • Two out of 4 of the original authors received funding from the Office of Director of National Intelligence and IARPA for their 2016 LFW survey follow up report
      • +
    • The LFW dataset includes over 500 actors, 30 models, 10 presidents, 24 football players, 124 basketball players, 11 kings, and 2 queens
      • +
    • In all the LFW publications provided by the authors the words "ethics", "consent", and "privacy" appear 0 times [^lfw_original_paper], [^lfw_survey], [^lfw_tech_report] , [^lfw_website]
      • +
    • The word "future" appears 71 times
      • +
    +

    Facts

    +
      +
    • Was created for the purpose of improving "unconstrained face recognition" [^lfw_original_paper]
      • +
    • All images in LFW were obtained "in the wild" meaning without any consent from the subject or from the photographer
      • +
    • The faces were detected using the Viola-Jones haarcascade face detector [^lfw_website] [^lfw_survey]
      • +
    • Is considered the "most popular benchmark for face recognition" [^lfw_baidu]
      • +
    • Is "the most widely used evaluation set in the field of facial recognition" [^lfw_pingan]
      • +

    • Is used by several of the largest tech companies in the world including "Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong." [^lfw_pingan]

      +
      • +

    • All images were copied from Yahoo News between 2002 - 2004 [^lfw_original_paper]

      +
      • +
    • SenseTime, who has relied on LFW for benchmarking their facial recognition performance, is the leading provider of surveillance to the Chinese Government
      • +
    +
    former President George W. Bush
    former President George W. Bush
    +
    Colin Powell (236), Tony Blair (144), and Donald Rumsfeld (121)
    Colin Powell (236), Tony Blair (144), and Donald Rumsfeld (121)

    People and Companies using the LFW Dataset

    +

    This section describes who is using the dataset and for what purposes. It should include specific examples of people or companies with citations and screenshots. This section is followed up by the graph, the map, and then the supplementary material.

    +

    The LFW dataset is used by numerous companies for benchmarking algorithms and in some cases training. According to the benchmarking results page [^lfw_results] provided by the authors, over 2 dozen companies have contributed their benchmark results.

    +

    According to BiometricUpdate.com [^lfw_pingan], LFW is "the most widely used evaluation set in the field of facial recognition, LFW attracts a few dozen teams from around the globe including Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong."

    +

    According to researchers at the Baidu Research – Institute of Deep Learning "LFW has been the most popular evaluation benchmark for face recognition, and played a very important role in facilitating the face recognition society to improve algorithm. [^lfw_baidu]."

    +

    In addition to commercial use as an evaluation tool, alll of the faces in LFW dataset are prepackaged into a popular machine learning code framework called scikit-learn.

    +
    "PING AN Tech facial recognition receives high score in latest LFW test results"
    "PING AN Tech facial recognition receives high score in latest LFW test results"
    +
    "Face Recognition Performance in LFW benchmark"
    "Face Recognition Performance in LFW benchmark"
    +
    "The 1st place in face verification challenge, LFW"
    "The 1st place in face verification challenge, LFW"

    In benchmarking, companies use a dataset to evaluate their algorithms which are typically trained on other data. After training, researchers will use LFW as a benchmark to compare results with other algorithms.

    +

    For example, Baidu (est. net worth $13B) uses LFW to report results for their "Targeting Ultimate Accuracy: Face Recognition via Deep Embedding". According to the three Baidu researchers who produced the paper:

    +

    Citations

    +

    Overall, LFW has at least 116 citations from 11 countries.

    +

    Conclusion

    +

    The LFW face recognition training and evaluation dataset is a historically important face dataset as it was the first popular dataset to be created entirely from Internet images, paving the way for a global trend towards downloading anyone’s face from the Internet and adding it to a dataset. As will be evident with other datasets, LFW’s approach has now become the norm.

    +

    For all the 5,000 people in this datasets, their face is forever a part of facial recognition history. It would be impossible to remove anyone from the dataset because it is so ubiquitous. For their rest of the lives and forever after, these 5,000 people will continue to be used for training facial recognition surveillance.

    +

    Code

    +
    #!/usr/bin/python
+
+import numpy as np
+from sklearn.datasets import fetch_lfw_people
+import imageio
+import imutils
+
+# download LFW dataset (first run takes a while)
+lfw_people = fetch_lfw_people(min_faces_per_person=1, resize=1, color=True, funneled=False)
+
+# introspect dataset
+n_samples, h, w, c = lfw_people.images.shape
+print(f'{n_samples:,} images at {w}x{h} pixels')
+cols, rows = (176, 76)
+n_ims = cols * rows
+
+# build montages
+im_scale = 0.5
+ims = lfw_people.images[:n_ims]
+montages = imutils.build_montages(ims, (int(w * im_scale,   int(h * im_scale)), (cols, rows))
+montage = montages[0]
+
+# save full montage image
+imageio.imwrite('lfw_montage_full.png', montage)
+
+# make a smaller version
+montage_960 = imutils.resize(montage, width=960)
+imageio.imwrite('lfw_montage_960.jpg', montage_960)
+
    +
    +
    +
      +
      +
      + +
      + + + + + \ No newline at end of file diff --git a/site/public/datasets_v0/lfw/right-to-removal/index.html b/site/public/datasets_v0/lfw/right-to-removal/index.html new file mode 100644 index 00000000..5dc269b2 --- /dev/null +++ b/site/public/datasets_v0/lfw/right-to-removal/index.html @@ -0,0 +1,62 @@ + + + + MegaPixels + + + + + + + + + + + + +
      + + +
      MegaPixels
      +       +
      + Datasets + Research + About +
      +
      +
      + +

      Labeled Faces in the Wild

      +

      Right to Removal

      +

      If you are affected by disclosure of your identity in this dataset please do contact the authors. Many have stated that they are willing to remove images upon request. The authors of the LFW dataset provide the following email for inquiries:

      +

      You can use the following message to request removal from the dataset:

      +

      To: Gary Huang mailto:gbhuang@cs.umass.edu

      +

      Subject: Request for Removal from LFW Face Dataset

      +

      Dear [researcher name],

      +

      I am writing to you about the "Labeled Faces in The Wild Dataset". Recently I discovered that your dataset includes my identity and I no longer wish to be included in your dataset.

      +

      The dataset is being used thousands of companies around the world to improve facial recognition software including usage by governments for the purpose of law enforcement, national security, tracking consumers in retail environments, and tracking individuals through public spaces.

      +

      My name as it appears in your dataset is [your name]. Please remove all images from your dataset and inform your newsletter subscribers to likewise update their copies.

      +

      - [your name]

      +
      +
      + +
      + + + + + \ No newline at end of file diff --git a/site/public/datasets_v0/lfw/tables/index.html b/site/public/datasets_v0/lfw/tables/index.html new file mode 100644 index 00000000..dd460843 --- /dev/null +++ b/site/public/datasets_v0/lfw/tables/index.html @@ -0,0 +1,52 @@ + + + + MegaPixels + + + + + + + + + + + + +
      + + +
      MegaPixels
      +       +
      + Datasets + Research + About +
      +
      +
      + +

      Labeled Faces in the Wild

      +

      Tables

      +
      + +
      + + + + + \ No newline at end of file diff --git a/site/public/datasets_v0/vgg_face2/index.html b/site/public/datasets_v0/vgg_face2/index.html new file mode 100644 index 00000000..b7ba5a4c --- /dev/null +++ b/site/public/datasets_v0/vgg_face2/index.html @@ -0,0 +1,80 @@ + + + + MegaPixels + + + + + + + + + + + + +
      + + +
      MegaPixels
      +       +
      + Datasets + Research + About +
      +
      +
      + +

      VGG Faces2

      +
      Created
      2018
      Images
      3.3M
      People
      9,000
      Created From
      Scraping search engines
      Search available
      [Searchable](#)

      VGG Face2 is the updated version of the VGG Face dataset and now includes over 3.3M face images from over 9K people. The identities were selected by taking the top 500K identities in Google's Knowledge Graph of celebrities and then selecting only the names that yielded enough training images. The dataset was created in the UK but funded by Office of Director of National Intelligence in the United States.

      +

      VGG Face2 by the Numbers

      +
        +
      • 1,331 actresses, 139 presidents
        • +
      • 3 husbands and 16 wives
        • +
      • 2 snooker player
        • +
      • 1 guru
        • +
      • 1 pornographic actress
        • +
      • 3 computer programmer
        • +
      +

      Names and descriptions

      +
        +
      • The original VGGF2 name list has been updated with the results returned from Google Knowledge
        • +
      • Names with a similarity score greater than 0.75 where automatically updated. Scores computed using import difflib; seq = difflib.SequenceMatcher(a=a.lower(), b=b.lower()); score = seq.ratio()
        • +
      • The 97 names with a score of 0.75 or lower were manually reviewed and includes name changes validating using Wikipedia.org results for names such as "Bruce Jenner" to "Caitlyn Jenner", spousal last-name changes, and discretionary changes to improve search results such as combining nicknames with full name when appropriate, for example changing "Aleksandar Petrović" to "Aleksandar 'Aco' Petrović" and minor changes such as "Mohammad Ali" to "Muhammad Ali"
        • +
      • The 'Description' text was automatically added when the Knowledge Graph score was greater than 250
        • +
      +

      TODO

      +
        +
      • create name list, and populate with Knowledge graph information like LFW
        • +
      • make list of interesting number stats, by the numbers
        • +
      • make list of interesting important facts
        • +
      • write intro abstract
        • +
      • write analysis of usage
        • +
      • find examples, citations, and screenshots of useage
        • +
      • find list of companies using it for table
        • +
      • create montages of the dataset, like LFW
        • +
      • create right to removal information
        • +
      +
      + +
      + + + + + \ No newline at end of file diff --git a/site/public/index.html b/site/public/index.html index d2986084..5038c483 100644 --- a/site/public/index.html +++ b/site/public/index.html @@ -86,15 +86,63 @@
      + +
      + Asian Face Age Dataset +
      +       + + +
      + Annotated Facial Landmarks in The Wild +
      +       + + +
      + Caltech 10K Faces Dataset +
      +       + + +
      + Caltech Occluded Faces in The Wild +
      +       + + +
      + FERET: FacE REcognition +
      +       + + +
      + Labeled Face Parts in The Wild +
      +       +
      Labeled Faces in The Wild
      + +
      + Unconstrained College Students +
      +       +
      - VGG Face2 + VGG Face 2 Dataset +
      +       + + +
      + YouTube Celebrities
      diff --git a/site/public/info/index.html b/site/public/info/index.html index d3a7d549..65510255 100644 --- a/site/public/info/index.html +++ b/site/public/info/index.html @@ -27,7 +27,7 @@
      -

      What do facial recognition algorithms see?

      +

      Results are only stored for the duration of the analysis and are deleted when you leave this page.

      diff --git a/site/public/research/00_introduction/index.html b/site/public/research/00_introduction/index.html index b6cc8e4a..64047134 100644 --- a/site/public/research/00_introduction/index.html +++ b/site/public/research/00_introduction/index.html @@ -42,18 +42,15 @@
      -
      Posted
      Dec. 15
      Author
      Adam Harvey

      It was the early 2000s. Face recognition was new and no one seemed sure exactly how well it was going to perform in practice. In theory, face recognition was poised to be a game changer, a force multiplier, a strategic military advantage, a way to make cities safer and to secure borders. This was the future John Ashcroft demanded with the Total Information Awareness act of the 2003 and that spooks had dreamed of for decades. It was a future that academics at Carnegie Mellon Universtiy and Colorado State University would help build. It was also a future that celebrities would play a significant role in building. And to the surprise of ordinary Internet users like myself and perhaps you, it was a future that millions of Internet users would unwittingly play role in creating.

      +
      Posted
      Dec. 15
      Author
      Adam Harvey

      Ignore content below these lines

      +

      Ever since the first computational facial recognition research project by the CIA in the early 1960s, data has always played a vital role in the development of our biometric future. Without facial recognition datasets there would be no facial recognition. Datasets are an indispensable part of any artificial intelligence system because, as Geoffrey Hinton points out, "we no longer program computers with code, we program them with data".

      +

      It was the early 2000s. Face recognition was new and no one seemed sure exactly how well it was going to perform in practice. In theory, face recognition was poised to be a game changer, a force multiplier, a strategic military advantage, a way to make cities safer and to secure borders. This was the future John Ashcroft demanded with the Total Information Awareness act of the 2003 and that spooks had dreamed of for decades. It was a future that academics at Carnegie Mellon Universtiy and Colorado State University would help build. It was also a future that celebrities would play a significant role in building. And to the surprise of ordinary Internet users like myself and perhaps you, it was a future that millions of Internet users would unwittingly play role in creating.

      Now the future has arrived and it doesn't make sense. Facial recognition works yet it doesn't actually work. Facial recognition is cheap and accessible but also expensive and out of control. Facial recognition research has achieved headline grabbing superhuman accuracies over 99.9% yet facial recognition is also dangerously inaccurate. During a trial installation at Sudkreuz station in Berlin in 2018, 20% of the matches were wrong, a number so low that it should not have any connection to law enforcement or justice. And in London, the Metropolitan police had been using facial recognition software that mistakenly identified an alarming 98% of people as criminals 1, which perhaps is a crime itself.

      MegaPixels is an online art project that explores the history of facial recognition from the perspective of datasets. To paraphrase the artist Trevor Paglen, whoever controls the dataset controls the meaning. MegaPixels aims to unravel the meanings behind the data and expose the darker corners of the biometric industry that have contributed to its growth. MegaPixels does not start with a conclusion, a moralistic slant, or a

      Whether or not to build facial recognition was a question that can no longer be asked. As an outspoken critic of face recognition I've developed, and hopefully furthered, my understanding during the last 10 years I've spent working with computer vision. Though I initially disagreed, I've come to see technocratic perspective as a non-negotiable reality. As Oren (nytimes article) wrote in NYT Op-Ed "the horse is out of the barn" and the only thing we can do collectively or individually is to steer towards the least worse outcome. Computational communication has entered a new era and it's both exciting and frightening to explore the potentials and opportunities. In 1997 getting access to 1 teraFLOPS of computational power would have cost you $55 million and required a strategic partnership with the Department of Defense. At the time of writing, anyone can rent 1 teraFLOPS on a cloud GPU marketplace for less than $1/day. 2.

      I hope that this project will illuminate the darker areas of strange world of facial recognition that have not yet received attention and encourage discourse in academic, industry, and . By no means do I believe discourse can save the day. Nor do I think creating artwork can. In fact, I'm not exactly sure what the outcome of this project will be. The project is not so much what I publish here but what happens after. This entire project is only a prologue.

      As McLuhan wrote, "You can't have a static, fixed position in the electric age". And in our hyper-connected age of mass surveillance, artificial intelligece, and unevenly distributed virtual futures the most irrational thing to be is rational. Increasingly the world is becoming a contradiction where people use surveillance to protest surveillance, use

      Like many projects, MegaPixels had spent years meandering between formats, unfeasible budgets, and was generally too niche of a subject. The basic idea for this project, as proposed to the original Glass Room installation in 2016 in NYC, was to build an interactive mirror that showed people if they had been included in the LFW facial recognition dataset. The idea was based on my reaction to all the datasets I'd come across during research for the CV Dazzle project. I'd noticed strange datasets created for training and testing face detection algorithms. Most were created in labratory settings and their interpretation of face data was very strict.

      -

      About the name

      -

      About the funding

      -

      About me

      -

      About the team

      -

      Conclusion

      for other post

      It was the early 2000s. Face recognition was new and no one seemed sure how well it was going to perform in practice. In theory, face recognition was poised to be a game changer, a force multiplier, a strategic military advantage, a way to make cities safer and to secure the borders. It was the future that John Ashcroft demanded with the Total Information Awareness act of the 2003. It was a future that academics helped build. It was a future that celebrities helped build. And it was a future that

      A decade earlier the Department of Homeland Security and the Counterdrug Technology Development Program Office initated a feasibilty study called FERET (FacE REcognition Technology) to "develop automatic face recognition capabilities that could be employed to assist security, intelligence, and law enforcement personnel in the performance of their duties [^feret_website]."

      -- cgit v1.2.3-70-g09d2 From 67896d3cdde877de940a282bebacd10ca1c56499 Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Wed, 27 Feb 2019 20:29:08 +0100 Subject: site watcher / loader --- README.md | 2 +- megapixels/app/site/builder.py | 22 ++-- megapixels/app/site/loader.py | 123 +++++++++++++++++++ megapixels/app/site/parser.py | 204 ++++++++----------------------- megapixels/commands/site/watch.py | 44 +++++++ site/assets/css/css.css | 1 + site/content/pages/datasets/lfw/index.md | 55 ++++----- site/public/datasets/lfw/index.html | 43 ++----- 8 files changed, 266 insertions(+), 228 deletions(-) create mode 100644 megapixels/app/site/loader.py create mode 100644 megapixels/commands/site/watch.py (limited to 'site') diff --git a/README.md b/README.md index e1a2c1d0..e46a6289 100644 --- a/README.md +++ b/README.md @@ -19,7 +19,7 @@ pip install numpy Pillow pip install dlib pip install requests simplejson click pdfminer.six pip install urllib3 flask flask_sqlalchemy mysql-connector -pip install pymediainfo tqdm opencv-python imutils +pip install pymediainfo tqdm opencv-python imutils watchdog pip install scikit-image python-dotenv imagehash scikit-learn colorlog pip install celery keras tensorflow pip install python.app # OSX only! needed for matplotlib diff --git a/megapixels/app/site/builder.py b/megapixels/app/site/builder.py index 188fbc25..15055110 100644 --- a/megapixels/app/site/builder.py +++ b/megapixels/app/site/builder.py @@ -7,6 +7,7 @@ from jinja2 import Environment, FileSystemLoader, select_autoescape import app.settings.app_cfg as cfg import app.site.s3 as s3 +import app.site.loader as loader import app.site.parser as parser env = Environment( @@ -21,7 +22,7 @@ def build_page(fn, research_posts, datasets): - syncs any assets with s3 - handles certain index pages... """ - metadata, sections = parser.read_metadata(fn) + metadata, sections = loader.read_metadata(fn) if metadata is None: print("{} has no metadata".format(fn)) @@ -55,7 +56,7 @@ def build_page(fn, research_posts, datasets): if 'index.md' in fn: s3.sync_directory(dirname, s3_dir, metadata) - content = parser.parse_markdown(sections, s3_path, skip_h1=skip_h1) + content = parser.parse_markdown(metadata, sections, s3_path, skip_h1=skip_h1) html = template.render( metadata=metadata, @@ -73,11 +74,11 @@ def build_index(key, research_posts, datasets): """ build the index of research (blog) posts """ - metadata, sections = parser.read_metadata(os.path.join(cfg.DIR_SITE_CONTENT, key, 'index.md')) + metadata, sections = loader.read_metadata(os.path.join(cfg.DIR_SITE_CONTENT, key, 'index.md')) template = env.get_template("page.html") s3_path = s3.make_s3_path(cfg.S3_SITE_PATH, metadata['path']) - content = parser.parse_markdown(sections, s3_path, skip_h1=False) - content += parser.parse_research_index(research_posts) + content = parser.parse_markdown(metadata, sections, s3_path, skip_h1=False) + content += loader.parse_research_index(research_posts) html = template.render( metadata=metadata, content=content, @@ -93,8 +94,8 @@ def build_site(): """ build the site! =^) """ - research_posts = parser.read_research_post_index() - datasets = parser.read_datasets_index() + research_posts = loader.read_research_post_index() + datasets = loader.read_datasets_index() for fn in glob.iglob(os.path.join(cfg.DIR_SITE_CONTENT, "**/*.md"), recursive=True): build_page(fn, research_posts, datasets) build_index('research', research_posts, datasets) @@ -103,7 +104,8 @@ def build_file(fn): """ build just one page from a filename! =^) """ - research_posts = parser.read_research_post_index() - datasets = parser.read_datasets_index() - fn = os.path.join(cfg.DIR_SITE_CONTENT, fn) + research_posts = loader.read_research_post_index() + datasets = loader.read_datasets_index() + if cfg.DIR_SITE_CONTENT not in fn: + fn = os.path.join(cfg.DIR_SITE_CONTENT, fn) build_page(fn, research_posts, datasets) diff --git a/megapixels/app/site/loader.py b/megapixels/app/site/loader.py new file mode 100644 index 00000000..691efb25 --- /dev/null +++ b/megapixels/app/site/loader.py @@ -0,0 +1,123 @@ +import os +import re +import glob +import simplejson as json + +import app.settings.app_cfg as cfg + +def read_metadata(fn): + """ + Read in read a markdown file and extract the metadata + """ + with open(fn, "r") as file: + data = file.read() + data = data.replace("\n ", "\n") + if "\n" in data: + data = data.replace("\r", "") + else: + data = data.replace("\r", "\n") + sections = data.split("\n\n") + return parse_metadata(fn, sections) + + +default_metadata = { + 'status': 'published', + 'title': 'Untitled Page', + 'desc': '', + 'slug': '', + 'published': '2018-12-31', + 'updated': '2018-12-31', + 'authors': 'Adam Harvey', + 'sync': 'true', + 'tagline': '', +} + +def parse_metadata(fn, sections): + """ + parse the metadata headers in a markdown file + (everything before the second ---------) + also generates appropriate urls for this page :) + """ + found_meta = False + metadata = {} + valid_sections = [] + for section in sections: + if not found_meta and ': ' in section: + found_meta = True + parse_metadata_section(metadata, section) + continue + if '-----' in section: + continue + if found_meta: + valid_sections.append(section) + + if 'title' not in metadata: + print('warning: {} has no title'.format(fn)) + for key in default_metadata: + if key not in metadata: + metadata[key] = default_metadata[key] + + basedir = os.path.dirname(fn.replace(cfg.DIR_SITE_CONTENT, '')) + basename = os.path.basename(fn) + if basedir == '/': + metadata['path'] = '/' + metadata['url'] = '/' + elif basename == 'index.md': + metadata['path'] = basedir + '/' + metadata['url'] = metadata['path'] + else: + metadata['path'] = basedir + '/' + metadata['url'] = metadata['path'] + basename.replace('.md', '') + '/' + + if metadata['status'] == 'published|draft|private': + metadata['status'] = 'published' + + metadata['sync'] = metadata['sync'] != 'false' + + metadata['author_html'] = '
      '.join(metadata['authors'].split(',')) + + return metadata, valid_sections + +def parse_metadata_section(metadata, section): + """ + parse a metadata key: value pair + """ + for line in section.split("\n"): + if ': ' not in line: + continue + key, value = line.split(': ', 1) + metadata[key.lower()] = value + + +def read_research_post_index(): + """ + Generate an index of the research (blog) posts + """ + return read_post_index('research') + + +def read_datasets_index(): + """ + Generate an index of the datasets + """ + return read_post_index('datasets') + + +def read_post_index(basedir): + """ + Generate an index of posts + """ + posts = [] + for fn in sorted(glob.glob(os.path.join(cfg.DIR_SITE_CONTENT, basedir, '*/index.md'))): + metadata, valid_sections = read_metadata(fn) + if metadata is None or metadata['status'] == 'private' or metadata['status'] == 'draft': + continue + posts.append(metadata) + if not len(posts): + posts.append({ + 'title': 'Placeholder', + 'slug': 'placeholder', + 'date': 'Placeholder', + 'url': '/', + }) + return posts diff --git a/megapixels/app/site/parser.py b/megapixels/app/site/parser.py index d6705214..3792e6f1 100644 --- a/megapixels/app/site/parser.py +++ b/megapixels/app/site/parser.py @@ -10,6 +10,49 @@ import app.site.s3 as s3 renderer = mistune.Renderer(escape=False) markdown = mistune.Markdown(renderer=renderer) +def parse_markdown(metadata, sections, s3_path, skip_h1=False): + """ + parse page into sections, preprocess the markdown to handle our modifications + """ + groups = [] + current_group = [] + for section in sections: + if skip_h1 and section.startswith('# '): + continue + elif section.strip().startswith('```'): + groups.append(format_section(current_group, s3_path)) + current_group = [] + current_group.append(section) + if section.strip().endswith('```'): + groups.append(format_applet("\n\n".join(current_group), s3_path)) + current_group = [] + elif section.strip().endswith('```'): + current_group.append(section) + groups.append(format_applet("\n\n".join(current_group), s3_path)) + current_group = [] + elif section.startswith('+ '): + groups.append(format_section(current_group, s3_path)) + groups.append(format_metadata(section)) + current_group = [] + elif '![fullwidth:' in section: + groups.append(format_section(current_group, s3_path)) + groups.append(format_section([section], s3_path, type='fullwidth')) + current_group = [] + elif '![wide:' in section: + groups.append(format_section(current_group, s3_path)) + groups.append(format_section([section], s3_path, type='wide')) + current_group = [] + elif '![' in section: + groups.append(format_section(current_group, s3_path)) + groups.append(format_section([section], s3_path, type='images')) + current_group = [] + else: + current_group.append(section) + groups.append(format_section(current_group, s3_path)) + content = "".join(groups) + return content + + def fix_images(lines, s3_path): """ do our own tranformation of the markdown around images to handle wide images etc @@ -32,6 +75,7 @@ def fix_images(lines, s3_path): real_lines.append(line) return "\n".join(real_lines) + def format_section(lines, s3_path, type=''): """ format a normal markdown section @@ -44,6 +88,7 @@ def format_section(lines, s3_path, type=''): return "
      " + markdown(lines) + "
      " return "" + def format_metadata(section): """ format a metadata section (+ key: value pairs) @@ -54,7 +99,11 @@ def format_metadata(section): meta.append("
      {}
      {}
      ".format(key, value)) return "
      {}
      ".format(''.join(meta)) + def format_applet(section, s3_path): + """ + Format the applets, which load javascript modules like the map and CSVs + """ # print(section) payload = section.strip('```').strip().strip('```').strip().split('\n') applet = {} @@ -79,47 +128,6 @@ def format_applet(section, s3_path): applet['fields'] = payload[1:] return "
      ".format(json.dumps(applet)) -def parse_markdown(sections, s3_path, skip_h1=False): - """ - parse page into sections, preprocess the markdown to handle our modifications - """ - groups = [] - current_group = [] - for section in sections: - if skip_h1 and section.startswith('# '): - continue - elif section.strip().startswith('```'): - groups.append(format_section(current_group, s3_path)) - current_group = [] - current_group.append(section) - if section.strip().endswith('```'): - groups.append(format_applet("\n\n".join(current_group), s3_path)) - current_group = [] - elif section.strip().endswith('```'): - current_group.append(section) - groups.append(format_applet("\n\n".join(current_group), s3_path)) - current_group = [] - elif section.startswith('+ '): - groups.append(format_section(current_group, s3_path)) - groups.append(format_metadata(section)) - current_group = [] - elif '![fullwidth:' in section: - groups.append(format_section(current_group, s3_path)) - groups.append(format_section([section], s3_path, type='fullwidth')) - current_group = [] - elif '![wide:' in section: - groups.append(format_section(current_group, s3_path)) - groups.append(format_section([section], s3_path, type='wide')) - current_group = [] - elif '![' in section: - groups.append(format_section(current_group, s3_path)) - groups.append(format_section([section], s3_path, type='images')) - current_group = [] - else: - current_group.append(section) - groups.append(format_section(current_group, s3_path)) - content = "".join(groups) - return content def parse_research_index(research_posts): """ @@ -141,117 +149,3 @@ def parse_research_index(research_posts): content += row content += '
      ' return content - -def read_metadata(fn): - """ - Read in read a markdown file and extract the metadata - """ - with open(fn, "r") as file: - data = file.read() - data = data.replace("\n ", "\n") - if "\n" in data: - data = data.replace("\r", "") - else: - data = data.replace("\r", "\n") - sections = data.split("\n\n") - return parse_metadata(fn, sections) - -default_metadata = { - 'status': 'published', - 'title': 'Untitled Page', - 'desc': '', - 'slug': '', - 'published': '2018-12-31', - 'updated': '2018-12-31', - 'authors': 'Adam Harvey', - 'sync': 'true', - 'tagline': '', -} - -def parse_metadata_section(metadata, section): - """ - parse a metadata key: value pair - """ - for line in section.split("\n"): - if ': ' not in line: - continue - key, value = line.split(': ', 1) - metadata[key.lower()] = value - -def parse_metadata(fn, sections): - """ - parse the metadata headers in a markdown file - (everything before the second ---------) - also generates appropriate urls for this page :) - """ - found_meta = False - metadata = {} - valid_sections = [] - for section in sections: - if not found_meta and ': ' in section: - found_meta = True - parse_metadata_section(metadata, section) - continue - if '-----' in section: - continue - if found_meta: - valid_sections.append(section) - - if 'title' not in metadata: - print('warning: {} has no title'.format(fn)) - for key in default_metadata: - if key not in metadata: - metadata[key] = default_metadata[key] - - basedir = os.path.dirname(fn.replace(cfg.DIR_SITE_CONTENT, '')) - basename = os.path.basename(fn) - if basedir == '/': - metadata['path'] = '/' - metadata['url'] = '/' - elif basename == 'index.md': - metadata['path'] = basedir + '/' - metadata['url'] = metadata['path'] - else: - metadata['path'] = basedir + '/' - metadata['url'] = metadata['path'] + basename.replace('.md', '') + '/' - - if metadata['status'] == 'published|draft|private': - metadata['status'] = 'published' - - metadata['sync'] = metadata['sync'] != 'false' - - metadata['author_html'] = '
      '.join(metadata['authors'].split(',')) - - return metadata, valid_sections - -def read_research_post_index(): - """ - Generate an index of the research (blog) posts - """ - return read_post_index('research') - -def read_datasets_index(): - """ - Generate an index of the datasets - """ - return read_post_index('datasets') - -def read_post_index(basedir): - """ - Generate an index of posts - """ - posts = [] - for fn in sorted(glob.glob(os.path.join(cfg.DIR_SITE_CONTENT, basedir, '*/index.md'))): - metadata, valid_sections = read_metadata(fn) - if metadata is None or metadata['status'] == 'private' or metadata['status'] == 'draft': - continue - posts.append(metadata) - if not len(posts): - posts.append({ - 'title': 'Placeholder', - 'slug': 'placeholder', - 'date': 'Placeholder', - 'url': '/', - }) - return posts - diff --git a/megapixels/commands/site/watch.py b/megapixels/commands/site/watch.py new file mode 100644 index 00000000..7fd3ba7c --- /dev/null +++ b/megapixels/commands/site/watch.py @@ -0,0 +1,44 @@ +""" +Watch for changes in the static site and build them +""" + +import click +import time +from watchdog.observers import Observer +from watchdog.events import PatternMatchingEventHandler + +import app.settings.app_cfg as cfg +from app.site.builder import build_site, build_file + +class SiteBuilder(PatternMatchingEventHandler): + """ + Handler for filesystem changes to the content path + """ + patterns = ["*.md"] + + def on_modified(self, event): + print(event.src_path, event.event_type) + build_file(event.src_path) + + def on_created(self, event): + print(event.src_path, event.event_type) + build_file(event.src_path) + +@click.command() +@click.pass_context +def cli(ctx): + """ + Run the observer and start watching for changes + """ + print("{} is now being watched for changes.".format(cfg.DIR_SITE_CONTENT)) + observer = Observer() + observer.schedule(SiteBuilder(), path=cfg.DIR_SITE_CONTENT, recursive=True) + observer.start() + + try: + while True: + time.sleep(1) + except KeyboardInterrupt: + observer.stop() + + observer.join() diff --git a/site/assets/css/css.css b/site/assets/css/css.css index 858d98eb..7b2e19fc 100644 --- a/site/assets/css/css.css +++ b/site/assets/css/css.css @@ -346,6 +346,7 @@ section.wide .image { } section.fullwidth { width: 100%; + background-size: contain; } section.fullwidth .image { max-width: 100%; diff --git a/site/content/pages/datasets/lfw/index.md b/site/content/pages/datasets/lfw/index.md index 8b37f035..48d86e1f 100644 --- a/site/content/pages/datasets/lfw/index.md +++ b/site/content/pages/datasets/lfw/index.md @@ -4,6 +4,8 @@ status: published title: Labeled Faces in The Wild desc: Labeled Faces in The Wild (LFW) is a database of face photographs designed for studying the problem of unconstrained face recognition subdesc: It includes 13,456 images of 4,432 people’s images copied from the Internet during 2002-2004. +image: lfw_index.gif +caption: Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms. slug: lfw published: 2019-2-23 updated: 2019-2-23 @@ -12,22 +14,13 @@ authors: Adam Harvey ------------ -# LFW +### Statistics + Years: 2002-2004 + Images: 13,233 + Identities: 5,749 + Origin: Yahoo News Images -+ Funding: (Possibly, partially CIA*) - -![fullwidth:Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.](assets/lfw_index.gif) - -*Labeled Faces in The Wild* (LFW) is "a database of face photographs designed for studying the problem of unconstrained face recognition[^lfw_www]. It is used to evaluate and improve the performance of facial recognition algorithms in academic, commercial, and government research. According to BiometricUpdate.com[^lfw_pingan], LFW is "the most widely used evaluation set in the field of facial recognition, LFW attracts a few dozen teams from around the globe including Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong." - -The LFW dataset includes 13,233 images of 5,749 people that were collected between 2002-2004. LFW is a subset of *Names of Faces* and is part of the first facial recognition training dataset created entirely from images appearing on the Internet. The people appearing in LFW are... - -The *Names and Faces* dataset was the first face recognition dataset created entire from online photos. However, *Names and Faces* and *LFW* are not the first face recognition dataset created entirely "in the wild". That title belongs to the [UCD dataset](/datasets/ucd_faces/). Images obtained "in the wild" means using an image without explicit consent or awareness from the subject or photographer. - ++ Funding: (Possibly, partially CIA) ### Analysis @@ -39,25 +32,35 @@ The *Names and Faces* dataset was the first face recognition dataset created ent - In all 3 of the LFW publications [^lfw_original_paper], [^lfw_survey], [^lfw_tech_report] the words "ethics", "consent", and "privacy" appear 0 times - The word "future" appears 71 times +## Labeled Faces in the Wild + +*Labeled Faces in The Wild* (LFW) is "a database of face photographs designed for studying the problem of unconstrained face recognition[^lfw_www]. It is used to evaluate and improve the performance of facial recognition algorithms in academic, commercial, and government research. According to BiometricUpdate.com[^lfw_pingan], LFW is "the most widely used evaluation set in the field of facial recognition, LFW attracts a few dozen teams from around the globe including Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong." + +The LFW dataset includes 13,233 images of 5,749 people that were collected between 2002-2004. LFW is a subset of *Names of Faces* and is part of the first facial recognition training dataset created entirely from images appearing on the Internet. The people appearing in LFW are... + +The *Names and Faces* dataset was the first face recognition dataset created entire from online photos. However, *Names and Faces* and *LFW* are not the first face recognition dataset created entirely "in the wild". That title belongs to the [UCD dataset](/datasets/ucd_faces/). Images obtained "in the wild" means using an image without explicit consent or awareness from the subject or photographer. + ### Synthetic Faces To visualize the types of photos in the dataset without explicitly publishing individual's identities a generative adversarial network (GAN) was trained on the entire dataset. The images in this video show a neural network learning the visual latent space and then interpolating between archetypical identities within the LFW dataset. ![fullwidth:](assets/lfw_synthetic.jpg) - ### Biometric Trade Routes To understand how this dataset has been used, its citations have been geocoded to show an approximate geographic digital trade route of the biometric data. Lines indicate an organization (education, commercial, or governmental) that has cited the LFW dataset in their research. Data is compiled from [SemanticScholar](https://www.semanticscholar.org). -[add map here] +``` +map +``` ### Citations Browse or download the geocoded citation data collected for the LFW dataset. -[add citations table here] - +``` +citations +``` ### Additional Information @@ -69,27 +72,14 @@ Browse or download the geocoded citation data collected for the LFW dataset. - The faces in the LFW dataset were detected using the Viola-Jones haarcascade face detector [^lfw_website] [^lfw-survey] - The LFW dataset is used by several of the largest tech companies in the world including "Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong." [^lfw_pingan] - All images in the LFW dataset were copied from Yahoo News between 2002 - 2004 -<<<<<<< HEAD -- In 2014, two of the four original authors of the LFW dataset received funding from IARPA and ODNI for their follow up paper [Labeled Faces in the Wild: Updates and New Reporting Procedures](https://www.semanticscholar.org/paper/Labeled-Faces-in-the-Wild-%3A-Updates-and-New-Huang-Learned-Miller/2d3482dcff69c7417c7b933f22de606a0e8e42d4) via IARPA contract number 2014-14071600010 +- In 2014, two of the four original authors of the LFW dataset received funding from IARPA and ODNI for their followup paper [Labeled Faces in the Wild: Updates and New Reporting Procedures](https://www.semanticscholar.org/paper/Labeled-Faces-in-the-Wild-%3A-Updates-and-New-Huang-Learned-Miller/2d3482dcff69c7417c7b933f22de606a0e8e42d4) via IARPA contract number 2014-14071600010 - The dataset includes 2 images of [George Tenet](http://vis-www.cs.umass.edu/lfw/person/George_Tenet.html), the former Director of Central Intelligence (DCI) for the Central Intelligence Agency whose facial biometrics were eventually used to help train facial recognition software in China and Russia -======= -- In 2014, 2/4 of the original authors of the LFW dataset received funding from IARPA and ODNI for their follow up paper "Labeled Faces in the Wild: Updates and New Reporting Procedures" via IARPA contract number 2014-14071600010 -- The LFW dataset was used Center for Intelligent Information Retrieval, the Central Intelligence Agency, the National Security Agency and National - -TODO (need citations for the following) - -- SenseTime, who has relied on LFW for benchmarking their facial recognition performance, is one the leading provider of surveillance to the Chinese Government [need citation for this fact. is it the most? or is that Tencent?] -- Two out of 4 of the original authors received funding from the Office of Director of National Intelligence and IARPA for their 2016 LFW survey follow up report - ->>>>>>> 13d7a450affe8ea4f368a97ea2014faa17702a4c ![Person with the most face images in LFW: former President George W. Bush](assets/lfw_montage_top1_640.jpg) ![Persons with the next most face images in LFW: Colin Powell (236), Tony Blair (144), and Donald Rumsfeld (121)](assets/lfw_montage_top2_4_640.jpg) ![All 5,379 faces in the Labeled Faces in The Wild Dataset](assets/lfw_montage_all_crop.jpg) - - ## Code The LFW dataset is so widely used that a popular code library called Sci-Kit Learn includes a function called `fetch_lfw_people` to download the faces in the LFW dataset. @@ -133,7 +123,6 @@ imageio.imwrite('lfw_montage_960.jpg', montage) ### Supplementary Material - ``` load_file assets/lfw_commercial_use.csv name_display, company_url, example_url, country, description @@ -141,14 +130,13 @@ name_display, company_url, example_url, country, description Text and graphics ©Adam Harvey / megapixels.cc - ------- Ignore text below these lines ------- -Research +### Research - "In our experiments, we used 10000 images and associated captions from the Faces in the wilddata set [3]." - "This work was supported in part by the Center for Intelligent Information Retrieval, the Central Intelligence Agency, the National Security Agency and National Science Foundation under CAREER award IIS-0546666 and grant IIS-0326249." @@ -159,6 +147,9 @@ Research - This research is based upon work supported in part by the Office of the Director of National Intelligence (ODNI), Intelligence Advanced Research Projects Activity (IARPA), via contract number 2014-14071600010. - From "Labeled Faces in the Wild: Updates and New Reporting Procedures" +### Footnotes + [^lfw_www]: [^lfw_baidu]: Jingtuo Liu, Yafeng Deng, Tao Bai, Zhengping Wei, Chang Huang. Targeting Ultimate Accuracy: Face Recognition via Deep Embedding. [^lfw_pingan]: Lee, Justin. "PING AN Tech facial recognition receives high score in latest LFW test results". BiometricUpdate.com. Feb 13, 2017. + diff --git a/site/public/datasets/lfw/index.html b/site/public/datasets/lfw/index.html index f83d8a66..86f49c52 100644 --- a/site/public/datasets/lfw/index.html +++ b/site/public/datasets/lfw/index.html @@ -27,11 +27,8 @@
      -

      LFW

      -
      Years
      2002-2004
      Images
      13,233
      Identities
      5,749
      Origin
      Yahoo News Images
      Funding
      (Possibly, partially CIA*)
      Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.
      Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.

      Labeled Faces in The Wild (LFW) is "a database of face photographs designed for studying the problem of unconstrained face recognition[^lfw_www]. It is used to evaluate and improve the performance of facial recognition algorithms in academic, commercial, and government research. According to BiometricUpdate.com[^lfw_pingan], LFW is "the most widely used evaluation set in the field of facial recognition, LFW attracts a few dozen teams from around the globe including Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong."

      -

      The LFW dataset includes 13,233 images of 5,749 people that were collected between 2002-2004. LFW is a subset of Names of Faces and is part of the first facial recognition training dataset created entirely from images appearing on the Internet. The people appearing in LFW are...

      -

      The Names and Faces dataset was the first face recognition dataset created entire from online photos. However, Names and Faces and LFW are not the first face recognition dataset created entirely "in the wild". That title belongs to the UCD dataset. Images obtained "in the wild" means using an image without explicit consent or awareness from the subject or photographer.

      -

      Analysis

      +

      Statistics

      +
      Years
      2002-2004
      Images
      13,233
      Identities
      5,749
      Origin
      Yahoo News Images
      Funding
      (Possibly, partially CIA)

      Analysis

      • There are about 3 men for every 1 woman (4,277 men and 1,472 women) in the LFW dataset[^lfw_www]
      • The person with the most images is George W. Bush with 530
        • @@ -41,15 +38,17 @@
      • In all 3 of the LFW publications [^lfw_original_paper], [^lfw_survey], [^lfw_tech_report] the words "ethics", "consent", and "privacy" appear 0 times
      • The word "future" appears 71 times
      +

      Labeled Faces in the Wild

      +

      Labeled Faces in The Wild (LFW) is "a database of face photographs designed for studying the problem of unconstrained face recognition[^lfw_www]. It is used to evaluate and improve the performance of facial recognition algorithms in academic, commercial, and government research. According to BiometricUpdate.com[^lfw_pingan], LFW is "the most widely used evaluation set in the field of facial recognition, LFW attracts a few dozen teams from around the globe including Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong."

      +

      The LFW dataset includes 13,233 images of 5,749 people that were collected between 2002-2004. LFW is a subset of Names of Faces and is part of the first facial recognition training dataset created entirely from images appearing on the Internet. The people appearing in LFW are...

      +

      The Names and Faces dataset was the first face recognition dataset created entire from online photos. However, Names and Faces and LFW are not the first face recognition dataset created entirely "in the wild". That title belongs to the UCD dataset. Images obtained "in the wild" means using an image without explicit consent or awareness from the subject or photographer.

      Synthetic Faces

      To visualize the types of photos in the dataset without explicitly publishing individual's identities a generative adversarial network (GAN) was trained on the entire dataset. The images in this video show a neural network learning the visual latent space and then interpolating between archetypical identities within the LFW dataset.

      Biometric Trade Routes

      To understand how this dataset has been used, its citations have been geocoded to show an approximate geographic digital trade route of the biometric data. Lines indicate an organization (education, commercial, or governmental) that has cited the LFW dataset in their research. Data is compiled from SemanticScholar.

      -

      [add map here]

      -

      Citations

      +

      Citations

      Browse or download the geocoded citation data collected for the LFW dataset.

      -

      [add citations table here]

      -

      Additional Information

      +

      Additional Information

      (tweet-sized snippets go here)

      • The LFW dataset is considered the "most popular benchmark for face recognition" [^lfw_baidu]
        • @@ -57,27 +56,10 @@
      • All images in LFW dataset were obtained "in the wild" meaning without any consent from the subject or from the photographer
      • The faces in the LFW dataset were detected using the Viola-Jones haarcascade face detector [^lfw_website] [^lfw-survey]
      • The LFW dataset is used by several of the largest tech companies in the world including "Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong." [^lfw_pingan]
        • -
      • All images in the LFW dataset were copied from Yahoo News between 2002 - 2004 -<<<<<<< HEAD
        • -
      • In 2014, two of the four original authors of the LFW dataset received funding from IARPA and ODNI for their follow up paper Labeled Faces in the Wild: Updates and New Reporting Procedures via IARPA contract number 2014-14071600010
        • -

      • The dataset includes 2 images of George Tenet, the former Director of Central Intelligence (DCI) for the Central Intelligence Agency whose facial biometrics were eventually used to help train facial recognition software in China and Russia

        -
        • -
      • In 2014, 2/4 of the original authors of the LFW dataset received funding from IARPA and ODNI for their follow up paper "Labeled Faces in the Wild: Updates and New Reporting Procedures" via IARPA contract number 2014-14071600010
        • -
      • The LFW dataset was used Center for Intelligent Information Retrieval, the Central Intelligence Agency, the National Security Agency and National
        • -
      -

      TODO (need citations for the following)

      -
        -
      • SenseTime, who has relied on LFW for benchmarking their facial recognition performance, is one the leading provider of surveillance to the Chinese Government [need citation for this fact. is it the most? or is that Tencent?]
        • -
      • Two out of 4 of the original authors received funding from the Office of Director of National Intelligence and IARPA for their 2016 LFW survey follow up report
        • +
      • All images in the LFW dataset were copied from Yahoo News between 2002 - 2004
        • +
      • In 2014, two of the four original authors of the LFW dataset received funding from IARPA and ODNI for their followup paper Labeled Faces in the Wild: Updates and New Reporting Procedures via IARPA contract number 2014-14071600010
        • +
      • The dataset includes 2 images of George Tenet, the former Director of Central Intelligence (DCI) for the Central Intelligence Agency whose facial biometrics were eventually used to help train facial recognition software in China and Russia
      -

      > 13d7a450affe8ea4f368a97ea2014faa17702a4c

      -
      -
      -
      -
      -
      -
      -
      former President George W. Bush
      former President George W. Bush
      Colin Powell (236), Tony Blair (144), and Donald Rumsfeld (121)
      Colin Powell (236), Tony Blair (144), and Donald Rumsfeld (121)
      All 5,379 faces in the Labeled Faces in The Wild Dataset
      All 5,379 faces in the Labeled Faces in The Wild Dataset

      Code

      The LFW dataset is so widely used that a popular code library called Sci-Kit Learn includes a function called fetch_lfw_people to download the faces in the LFW dataset.

      @@ -113,7 +95,7 @@ imageio.imwrite('lfw_montage_960.jpg', montage)

      Supplementary Material

      Text and graphics ©Adam Harvey / megapixels.cc

      Ignore text below these lines

      -

      Research

      +

      Research

      • "In our experiments, we used 10000 images and associated captions from the Faces in the wilddata set [3]."
      • "This work was supported in part by the Center for Intelligent Information Retrieval, the Central Intelligence Agency, the National Security Agency and National Science Foundation under CAREER award IIS-0546666 and grant IIS-0326249."
        • @@ -125,6 +107,7 @@ imageio.imwrite('lfw_montage_960.jpg', montage)
      • From "Labeled Faces in the Wild: Updates and New Reporting Procedures"
      +

      Footnotes

        -- cgit v1.2.3-70-g09d2 From 9bac173e85865e4f0d1dba5071b40eb7ebe3dd1a Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Wed, 27 Feb 2019 22:15:03 +0100 Subject: new intro header for datasets page and sidebar --- client/index.js | 6 +-- megapixels/app/site/parser.py | 70 ++++++++++++++++++++++++++---- megapixels/commands/site/watch.py | 2 + site/assets/css/css.css | 72 ++++++++++++++++++++++++++----- site/assets/css/tabulator.css | 2 +- site/content/pages/datasets/lfw/index.md | 25 +++++------ site/content/pages/datasets/uccs/index.md | 2 +- site/public/datasets/lfw/index.html | 36 ++++------------ 8 files changed, 152 insertions(+), 63 deletions(-) (limited to 'site') diff --git a/client/index.js b/client/index.js index c9335f14..37906f30 100644 --- a/client/index.js +++ b/client/index.js @@ -110,9 +110,9 @@ function runApplets() { function main() { const paras = document.querySelectorAll('section p') - if (paras.length) { - paras[0].classList.add('first_paragraph') - } + // if (paras.length) { + // paras[0].classList.add('first_paragraph') + // } toArray(document.querySelectorAll('header .links a')).forEach(tag => { if (window.location.href.match(tag.href)) { tag.classList.add('active') diff --git a/megapixels/app/site/parser.py b/megapixels/app/site/parser.py index 3792e6f1..dc53177b 100644 --- a/megapixels/app/site/parser.py +++ b/megapixels/app/site/parser.py @@ -16,9 +16,30 @@ def parse_markdown(metadata, sections, s3_path, skip_h1=False): """ groups = [] current_group = [] + in_stats = False + + if 'desc' in metadata and 'subdesc' in metadata: + groups.append(intro_section(metadata, s3_path)) + for section in sections: if skip_h1 and section.startswith('# '): continue + elif section.strip().startswith('---'): + continue + elif section.lower().strip().startswith('ignore text'): + break + elif '### Statistics' in section: + if len(current_group): + groups.append(format_section(current_group, s3_path)) + current_group = [] + current_group.append(section) + in_stats = True + elif in_stats and not section.strip().startswith('## '): + current_group.append(section) + elif in_stats and section.strip().startswith('## '): + current_group = [format_section(current_group, s3_path, 'right-sidebar', tag='div')] + current_group.append(section) + in_stats = False elif section.strip().startswith('```'): groups.append(format_section(current_group, s3_path)) current_group = [] @@ -32,7 +53,7 @@ def parse_markdown(metadata, sections, s3_path, skip_h1=False): current_group = [] elif section.startswith('+ '): groups.append(format_section(current_group, s3_path)) - groups.append(format_metadata(section)) + groups.append('
        ' + format_metadata(section) + '
        ') current_group = [] elif '![fullwidth:' in section: groups.append(format_section(current_group, s3_path)) @@ -52,6 +73,32 @@ def parse_markdown(metadata, sections, s3_path, skip_h1=False): content = "".join(groups) return content +def intro_section(metadata, s3_path): + """ + Build the intro section for datasets + """ + + section = "
        ".format(s3_path + metadata['image']) + section += "
        " + + parts = [] + if 'desc' in metadata: + desc = metadata['desc'] + if 'color' in metadata and metadata['title'] in desc: + desc = desc.replace(metadata['title'], "{}".format(metadata['color'], metadata['title'])) + section += "
        {}
        ".format(desc, desc) + + if 'subdesc' in metadata: + subdesc = markdown(metadata['subdesc']).replace('

        ', '').replace('

        ', '') + section += "
        {}
        ".format(subdesc, subdesc) + + section += "
        " + section += "
        " + + if 'caption' in metadata: + section += "
        {}
        ".format(metadata['caption']) + + return section def fix_images(lines, s3_path): """ @@ -75,19 +122,26 @@ def fix_images(lines, s3_path): real_lines.append(line) return "\n".join(real_lines) - -def format_section(lines, s3_path, type=''): +def format_section(lines, s3_path, type='', tag='section'): """ format a normal markdown section """ if len(lines): + lines = fix_meta(lines) lines = fix_images(lines, s3_path) if type: - return "
        {}
        ".format(type, markdown(lines)) + return "<{} class='{}'>{}".format(tag, type, markdown(lines), tag) else: - return "
        " + markdown(lines) + "
        " + return "<{}>{}".format(tag, markdown(lines), tag) return "" +def fix_meta(lines): + new_lines = [] + for line in lines: + if line.startswith('+ '): + line = format_metadata(line) + new_lines.append(line) + return new_lines def format_metadata(section): """ @@ -97,8 +151,7 @@ def format_metadata(section): for line in section.split('\n'): key, value = line[2:].split(': ', 1) meta.append("
        {}
        {}
        ".format(key, value)) - return "
        {}
        ".format(''.join(meta)) - + return "
        {}
        ".format(''.join(meta)) def format_applet(section, s3_path): """ @@ -107,12 +160,13 @@ def format_applet(section, s3_path): # print(section) payload = section.strip('```').strip().strip('```').strip().split('\n') applet = {} - print(payload) + # print(payload) if ': ' in payload[0]: command, opt = payload[0].split(': ') else: command = payload[0] opt = None + print(command) if command == 'python' or command == 'javascript' or command == 'code': return format_section([ section ], s3_path) if command == '': diff --git a/megapixels/commands/site/watch.py b/megapixels/commands/site/watch.py index 7fd3ba7c..7bd71038 100644 --- a/megapixels/commands/site/watch.py +++ b/megapixels/commands/site/watch.py @@ -35,6 +35,8 @@ def cli(ctx): observer.schedule(SiteBuilder(), path=cfg.DIR_SITE_CONTENT, recursive=True) observer.start() + build_file(cfg.DIR_SITE_CONTENT + "/datasets/lfw/index.md") + try: while True: time.sleep(1) diff --git a/site/assets/css/css.css b/site/assets/css/css.css index 7b2e19fc..fed381a7 100644 --- a/site/assets/css/css.css +++ b/site/assets/css/css.css @@ -4,12 +4,12 @@ html, body { padding: 0; width: 100%; min-height: 100%; - font-family: 'Roboto', sans-serif; - color: #b8b8b8; + font-family: 'Roboto Mono', sans-serif; + color: #eee; overflow-x: hidden; } html { - background: #191919; + background: #111111; } .content { @@ -146,8 +146,8 @@ h2 { h3 { margin: 0 0 20px 0; padding: 0; - font-size: 11pt; - font-weight: 500; + font-size: 14pt; + font-weight: 600; transition: color 0.2s cubic-bezier(0,0,1,1); } h4 { @@ -165,8 +165,15 @@ h4 { color: #fff; text-decoration: underline; } +.right-sidebar h3 { + margin: 0; + padding: 0 0 10px 0; + font-family: 'Roboto Mono'; + text-transform: uppercase; + letter-spacing: 2px; +} -th, .gray, h3, h4 { +th, .gray { font-family: 'Roboto Mono', monospace; font-weight: 400; text-transform: uppercase; @@ -201,6 +208,7 @@ section { } p { margin: 0 0 20px 0; + line-height: 2; } .content a { color: #ddd; @@ -229,10 +237,13 @@ p { } .right-sidebar { float: right; - width: 200px; + width: 240px; margin-left: 20px; + padding-top: 10px; padding-left: 20px; border-left: 1px solid #444; + font-family: 'Roboto'; + font-size: 14px; } .right-sidebar .meta { flex-direction: column; @@ -240,6 +251,9 @@ p { .right-sidebar .meta > div { margin-bottom: 10px; } +.right-sidebar ul { + margin-bottom: 10px; +} /* lists */ @@ -346,17 +360,17 @@ section.wide .image { } section.fullwidth { width: 100%; - background-size: contain; } section.fullwidth .image { max-width: 100%; } .caption { - text-align: center; + text-align: left; font-size: 9pt; - color: #888; - max-width: 620px; + color: #bbb; + max-width: 960px; margin: 10px auto 0 auto; + font-family: 'Roboto'; } /* blog index */ @@ -499,3 +513,39 @@ section.fullwidth .image { .dataset-list a:nth-child(3n+3) { background-color: rgba(255, 255, 0, 0.1); } .desktop .dataset-list .dataset:nth-child(3n+3):hover { background-color: rgba(255, 255, 0, 0.2); } + + +/* intro section for datasets */ + +section.intro_section { + font-family: 'Roboto Mono'; + width: 100%; + background-size: cover; + background-position: bottom left; + padding: 50px 0; + min-height: 60vh; + display: flex; + justify-content: center; + align-items: center; + background-color: #111111; +} +.intro_section .inner { + max-width: 960px; + margin: 0 auto; +} +.intro_section .hero_desc { + font-size: 38px; + line-height: 60px; + margin-bottom: 30px; + color: #fff; +} +.intro_section .hero_subdesc { + font-size: 18px; + line-height: 36px; + max-width: 640px; + color: #ddd; +} +.intro_section span { + box-shadow: -10px -10px #000, 10px -10px #000, 10px 10px #000, -10px 10px #000; + background: #000; +} \ No newline at end of file diff --git a/site/assets/css/tabulator.css b/site/assets/css/tabulator.css index 200f0c5c..63abf050 100755 --- a/site/assets/css/tabulator.css +++ b/site/assets/css/tabulator.css @@ -493,7 +493,7 @@ display: inline-block; position: relative; box-sizing: border-box; - padding: 4px; + padding: 10px; border-right: 1px solid #333; vertical-align: middle; white-space: nowrap; diff --git a/site/content/pages/datasets/lfw/index.md b/site/content/pages/datasets/lfw/index.md index 48d86e1f..1995e1f9 100644 --- a/site/content/pages/datasets/lfw/index.md +++ b/site/content/pages/datasets/lfw/index.md @@ -2,14 +2,14 @@ status: published title: Labeled Faces in The Wild -desc: Labeled Faces in The Wild (LFW) is a database of face photographs designed for studying the problem of unconstrained face recognition +desc: Labeled Faces in The Wild (LFW) is a database of face photographs designed for studying the problem of unconstrained face recognition. subdesc: It includes 13,456 images of 4,432 people’s images copied from the Internet during 2002-2004. -image: lfw_index.gif +image: assets/lfw_feature.jpg caption: Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms. slug: lfw published: 2019-2-23 updated: 2019-2-23 -color: #00FF00 +color: #ff0000 authors: Adam Harvey ------------ @@ -22,12 +22,11 @@ authors: Adam Harvey + Origin: Yahoo News Images + Funding: (Possibly, partially CIA) -### Analysis +### INSIGHTS - There are about 3 men for every 1 woman (4,277 men and 1,472 women) in the LFW dataset[^lfw_www] - The person with the most images is [George W. Bush](http://vis-www.cs.umass.edu/lfw/person/George_W_Bush_comp.html) with 530 - There are about 3 George W. Bush's for every 1 [Tony Blair](http://vis-www.cs.umass.edu/lfw/person/Tony_Blair.html) -- 70% of people in the dataset have only 1 image and 29% have 2 or more images - The LFW dataset includes over 500 actors, 30 models, 10 presidents, 124 basketball players, 24 football players, 11 kings, 7 queens, and 1 [Moby](http://vis-www.cs.umass.edu/lfw/person/Moby.html) - In all 3 of the LFW publications [^lfw_original_paper], [^lfw_survey], [^lfw_tech_report] the words "ethics", "consent", and "privacy" appear 0 times - The word "future" appears 71 times @@ -40,20 +39,20 @@ The LFW dataset includes 13,233 images of 5,749 people that were collected betwe The *Names and Faces* dataset was the first face recognition dataset created entire from online photos. However, *Names and Faces* and *LFW* are not the first face recognition dataset created entirely "in the wild". That title belongs to the [UCD dataset](/datasets/ucd_faces/). Images obtained "in the wild" means using an image without explicit consent or awareness from the subject or photographer. -### Synthetic Faces - -To visualize the types of photos in the dataset without explicitly publishing individual's identities a generative adversarial network (GAN) was trained on the entire dataset. The images in this video show a neural network learning the visual latent space and then interpolating between archetypical identities within the LFW dataset. - -![fullwidth:](assets/lfw_synthetic.jpg) - ### Biometric Trade Routes -To understand how this dataset has been used, its citations have been geocoded to show an approximate geographic digital trade route of the biometric data. Lines indicate an organization (education, commercial, or governmental) that has cited the LFW dataset in their research. Data is compiled from [SemanticScholar](https://www.semanticscholar.org). +To understand how this dataset has been used, its citations have been geocoded to show an approximate geographic digital trade route of the biometric data. Lines indicate an organization (education, commercial, or governmental) that has cited the LFW dataset in their research. Data is compiled from [Semantic Scholar](https://www.semanticscholar.org). ``` map ``` +### Synthetic Faces + +To visualize the types of photos in the dataset without explicitly publishing individual's identities a generative adversarial network (GAN) was trained on the entire dataset. The images in this video show a neural network learning the visual latent space and then interpolating between archetypical identities within the LFW dataset. + +![fullwidth:](assets/lfw_synthetic.jpg) + ### Citations Browse or download the geocoded citation data collected for the LFW dataset. @@ -136,6 +135,7 @@ Ignore text below these lines ------- + ### Research - "In our experiments, we used 10000 images and associated captions from the Faces in the wilddata set [3]." @@ -146,6 +146,7 @@ Ignore text below these lines - This research is based upon work supported in part by the Office of the Director of National Intelligence (ODNI), Intelligence Advanced Research Projects Activity (IARPA), via contract number 2014-14071600010. - From "Labeled Faces in the Wild: Updates and New Reporting Procedures" +- 70% of people in the dataset have only 1 image and 29% have 2 or more images ### Footnotes diff --git a/site/content/pages/datasets/uccs/index.md b/site/content/pages/datasets/uccs/index.md index d40dce22..be1d2474 100644 --- a/site/content/pages/datasets/uccs/index.md +++ b/site/content/pages/datasets/uccs/index.md @@ -68,7 +68,7 @@ The more recent UCCS version of the dataset received funding from [^funding_uccs - You are welcomed to use these images for academic and journalistic use including for research papers, news stories, presentations. - Please use the following citation: -```MegaPixels.cc Adam Harvey 2013-2109.``` +```MegaPixels.cc Adam Harvey 2013-2019.``` [^funding_sb]: Sapkota, Archana and Boult, Terrance. "Large Scale Unconstrained Open Set Face Database." 2013. [^funding_uccs]: Günther, M. et. al. "Unconstrained Face Detection and Open-Set Face Recognition Challenge," 2018. Arxiv 1708.02337v3. \ No newline at end of file diff --git a/site/public/datasets/lfw/index.html b/site/public/datasets/lfw/index.html index 86f49c52..1242df0c 100644 --- a/site/public/datasets/lfw/index.html +++ b/site/public/datasets/lfw/index.html @@ -4,7 +4,7 @@ MegaPixels - + @@ -27,26 +27,26 @@
        -

        Statistics

        -
        Years
        2002-2004
        Images
        13,233
        Identities
        5,749
        Origin
        Yahoo News Images
        Funding
        (Possibly, partially CIA)

        Analysis

        +
        Labeled Faces in The Wild (LFW) is a database of face photographs designed for studying the problem of unconstrained face recognition.
        It includes 13,456 images of 4,432 people’s images copied from the Internet during 2002-2004. +
        Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.

        Statistics

        +
        Years
        2002-2004
        Images
        13,233
        Identities
        5,749
        Origin
        Yahoo News Images
        Funding
        (Possibly, partially CIA)

        INSIGHTS

        • There are about 3 men for every 1 woman (4,277 men and 1,472 women) in the LFW dataset[^lfw_www]
        • The person with the most images is George W. Bush with 530
        • There are about 3 George W. Bush's for every 1 Tony Blair
          • -
        • 70% of people in the dataset have only 1 image and 29% have 2 or more images
        • The LFW dataset includes over 500 actors, 30 models, 10 presidents, 124 basketball players, 24 football players, 11 kings, 7 queens, and 1 Moby
        • In all 3 of the LFW publications [^lfw_original_paper], [^lfw_survey], [^lfw_tech_report] the words "ethics", "consent", and "privacy" appear 0 times
        • The word "future" appears 71 times
        -

        Labeled Faces in the Wild

        +

        Labeled Faces in the Wild

        Labeled Faces in The Wild (LFW) is "a database of face photographs designed for studying the problem of unconstrained face recognition[^lfw_www]. It is used to evaluate and improve the performance of facial recognition algorithms in academic, commercial, and government research. According to BiometricUpdate.com[^lfw_pingan], LFW is "the most widely used evaluation set in the field of facial recognition, LFW attracts a few dozen teams from around the globe including Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong."

        The LFW dataset includes 13,233 images of 5,749 people that were collected between 2002-2004. LFW is a subset of Names of Faces and is part of the first facial recognition training dataset created entirely from images appearing on the Internet. The people appearing in LFW are...

        The Names and Faces dataset was the first face recognition dataset created entire from online photos. However, Names and Faces and LFW are not the first face recognition dataset created entirely "in the wild". That title belongs to the UCD dataset. Images obtained "in the wild" means using an image without explicit consent or awareness from the subject or photographer.

        -

        Synthetic Faces

        +

        Biometric Trade Routes

        +

        To understand how this dataset has been used, its citations have been geocoded to show an approximate geographic digital trade route of the biometric data. Lines indicate an organization (education, commercial, or governmental) that has cited the LFW dataset in their research. Data is compiled from Semantic Scholar.

        +

        Synthetic Faces

        To visualize the types of photos in the dataset without explicitly publishing individual's identities a generative adversarial network (GAN) was trained on the entire dataset. The images in this video show a neural network learning the visual latent space and then interpolating between archetypical identities within the LFW dataset.

        -

        Biometric Trade Routes

        -

        To understand how this dataset has been used, its citations have been geocoded to show an approximate geographic digital trade route of the biometric data. Lines indicate an organization (education, commercial, or governmental) that has cited the LFW dataset in their research. Data is compiled from SemanticScholar.

        -

        Citations

        +

        Citations

        Browse or download the geocoded citation data collected for the LFW dataset.

        Additional Information

        (tweet-sized snippets go here)

        @@ -94,24 +94,6 @@ imageio.imwrite('lfw_montage_960.jpg', montage)

        Supplementary Material

        Text and graphics ©Adam Harvey / megapixels.cc

        -

        Ignore text below these lines

        -

        Research

        -
          -
        • "In our experiments, we used 10000 images and associated captions from the Faces in the wilddata set [3]."
          • -
        • "This work was supported in part by the Center for Intelligent Information Retrieval, the Central Intelligence Agency, the National Security Agency and National Science Foundation under CAREER award IIS-0546666 and grant IIS-0326249."
          • -
        • From: "People-LDA: Anchoring Topics to People using Face Recognition" https://www.semanticscholar.org/paper/People-LDA%3A-Anchoring-Topics-to-People-using-Face-Jain-Learned-Miller/10f17534dba06af1ddab96c4188a9c98a020a459 and https://ieeexplore.ieee.org/document/4409055
          • -
        • This paper was presented at IEEE 11th ICCV conference Oct 14-21 and the main LFW paper "Labeled Faces in the Wild: A Database for Studying Face Recognition in Unconstrained Environments" was also published that same year
          • -

        • 10f17534dba06af1ddab96c4188a9c98a020a459

          -
          • -

        • This research is based upon work supported in part by the Office of the Director of National Intelligence (ODNI), Intelligence Advanced Research Projects Activity (IARPA), via contract number 2014-14071600010.

          -
          • -
        • From "Labeled Faces in the Wild: Updates and New Reporting Procedures"
          • -
        -

        Footnotes

        -
        -
        -
          -
          -- cgit v1.2.3-70-g09d2 From 1b008e4b4d11def9b13dc0a800b0d068624d43ae Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Wed, 27 Feb 2019 23:48:35 +0100 Subject: half of a footnote implementation --- megapixels/app/site/parser.py | 35 +++++++++++++++++++++++++++++------ site/assets/css/css.css | 34 ++++++++++++++++++++++++++++++++++ site/public/datasets/lfw/index.html | 15 +++++++++------ 3 files changed, 72 insertions(+), 12 deletions(-) (limited to 'site') diff --git a/megapixels/app/site/parser.py b/megapixels/app/site/parser.py index 98d9f284..ef83b655 100644 --- a/megapixels/app/site/parser.py +++ b/megapixels/app/site/parser.py @@ -18,6 +18,7 @@ def parse_markdown(metadata, sections, s3_path, skip_h1=False): current_group = [] footnotes = [] in_stats = False + in_footnotes = False ignoring = False if 'desc' in metadata and 'subdesc' in metadata: @@ -33,6 +34,7 @@ def parse_markdown(metadata, sections, s3_path, skip_h1=False): continue elif section.strip().startswith('### Footnotes'): groups.append(format_section(current_group, s3_path)) + current_group = [] footnotes = [] in_footnotes = True elif in_footnotes: @@ -82,10 +84,18 @@ def parse_markdown(metadata, sections, s3_path, skip_h1=False): current_group.append(section) groups.append(format_section(current_group, s3_path)) + footnote_txt = '' + footnote_lookup = {} + if len(footnotes): - groups.append(format_footnotes(footnotes, s3_path)) + footnote_txt, footnote_lookup = format_footnotes(footnotes, s3_path) content = "".join(groups) + + if footnote_lookup: + for key, index in footnote_lookup.items(): + content = content.replace(key, '{}'.format(key, index, index)) + content += footnote_txt return content @@ -153,8 +163,10 @@ def format_section(lines, s3_path, type='', tag='section'): return "<{}>{}".format(tag, markdown(lines), tag) return "" - def fix_meta(lines): + """ + Format metadata sections before passing to markdown + """ new_lines = [] for line in lines: if line.startswith('+ '): @@ -162,7 +174,6 @@ def fix_meta(lines): new_lines.append(line) return new_lines - def format_metadata(section): """ format a metadata section (+ key: value pairs) @@ -173,12 +184,24 @@ def format_metadata(section): meta.append("
          {}
          {}
          ".format(key, value)) return "
          {}
          ".format(''.join(meta)) -def format_footnotes(footnotes): +def format_footnotes(footnotes, s3_path): + """ + Format the footnotes section separately and produce a lookup we can use to update the main site + """ footnotes = '\n'.join(footnotes).split('\n') + index = 1 + footnote_index_lookup = {} + footnote_list = [] for footnote in footnotes: if not len(footnote) or '[^' not in footnote: continue - key, footnote = footnotes.split(': ') + key, note = footnote.split(': ', 1) + footnote_index_lookup[key] = index + footnote_list.append('^'.format(key) + markdown(note)) + index += 1 + + footnote_txt = '
          • ' + '
          • '.join(footnote_list) + '
          ' + return footnote_txt, footnote_index_lookup def format_applet(section, s3_path): """ @@ -189,7 +212,7 @@ def format_applet(section, s3_path): applet = {} # print(payload) if ': ' in payload[0]: - command, opt = payload[0].split(': ') + command, opt = payload[0].split(': ', 1) else: command = payload[0] opt = None diff --git a/site/assets/css/css.css b/site/assets/css/css.css index fed381a7..8b4241ea 100644 --- a/site/assets/css/css.css +++ b/site/assets/css/css.css @@ -548,4 +548,38 @@ section.intro_section { .intro_section span { box-shadow: -10px -10px #000, 10px -10px #000, 10px 10px #000, -10px 10px #000; background: #000; +} + +/* footnotes */ + +a.footnote { + font-size: 10px; + position: relative; + display: inline-block; + bottom: 10px; + text-decoration: none; + color: #ff0; + left: 2px; +} +.right-sidebar a.footnote { + bottom: 8px; +} +.desktop a.footnote:hover { + background-color: #ff0; + color: #000; +} +a.footnote_anchor { + font-weight: bold; + color: #ff0; + margin-right: 10px; + text-decoration: underline; + cursor: pointer; +} +ul.footnotes { + list-style-type: decimal; + margin-left: 30px; +} +li p { + margin: 0; padding: 0; + display: inline; } \ No newline at end of file diff --git a/site/public/datasets/lfw/index.html b/site/public/datasets/lfw/index.html index 1242df0c..54b6aa22 100644 --- a/site/public/datasets/lfw/index.html +++ b/site/public/datasets/lfw/index.html @@ -31,7 +31,7 @@
          Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.

          Statistics

          Years
          2002-2004
          Images
          13,233
          Identities
          5,749
          Origin
          Yahoo News Images
          Funding
          (Possibly, partially CIA)

          INSIGHTS

            -
          • There are about 3 men for every 1 woman (4,277 men and 1,472 women) in the LFW dataset[^lfw_www]
            • +
          • There are about 3 men for every 1 woman (4,277 men and 1,472 women) in the LFW dataset1
          • The person with the most images is George W. Bush with 530
          • There are about 3 George W. Bush's for every 1 Tony Blair
          • The LFW dataset includes over 500 actors, 30 models, 10 presidents, 124 basketball players, 24 football players, 11 kings, 7 queens, and 1 Moby
            • @@ -39,7 +39,7 @@
          • The word "future" appears 71 times

          Labeled Faces in the Wild

          -

          Labeled Faces in The Wild (LFW) is "a database of face photographs designed for studying the problem of unconstrained face recognition[^lfw_www]. It is used to evaluate and improve the performance of facial recognition algorithms in academic, commercial, and government research. According to BiometricUpdate.com[^lfw_pingan], LFW is "the most widely used evaluation set in the field of facial recognition, LFW attracts a few dozen teams from around the globe including Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong."

          +

          Labeled Faces in The Wild (LFW) is "a database of face photographs designed for studying the problem of unconstrained face recognition1. It is used to evaluate and improve the performance of facial recognition algorithms in academic, commercial, and government research. According to BiometricUpdate.com3, LFW is "the most widely used evaluation set in the field of facial recognition, LFW attracts a few dozen teams from around the globe including Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong."

          The LFW dataset includes 13,233 images of 5,749 people that were collected between 2002-2004. LFW is a subset of Names of Faces and is part of the first facial recognition training dataset created entirely from images appearing on the Internet. The people appearing in LFW are...

          The Names and Faces dataset was the first face recognition dataset created entire from online photos. However, Names and Faces and LFW are not the first face recognition dataset created entirely "in the wild". That title belongs to the UCD dataset. Images obtained "in the wild" means using an image without explicit consent or awareness from the subject or photographer.

          Biometric Trade Routes

          @@ -51,11 +51,11 @@

          Additional Information

          (tweet-sized snippets go here)

            -
          • The LFW dataset is considered the "most popular benchmark for face recognition" [^lfw_baidu]
            • -
          • The LFW dataset is "the most widely used evaluation set in the field of facial recognition" [^lfw_pingan]
            • +
          • The LFW dataset is considered the "most popular benchmark for face recognition" 2
            • +
          • The LFW dataset is "the most widely used evaluation set in the field of facial recognition" 3
          • All images in LFW dataset were obtained "in the wild" meaning without any consent from the subject or from the photographer
          • The faces in the LFW dataset were detected using the Viola-Jones haarcascade face detector [^lfw_website] [^lfw-survey]
            • -
          • The LFW dataset is used by several of the largest tech companies in the world including "Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong." [^lfw_pingan]
            • +
          • The LFW dataset is used by several of the largest tech companies in the world including "Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong." 3
          • All images in the LFW dataset were copied from Yahoo News between 2002 - 2004
          • In 2014, two of the four original authors of the LFW dataset received funding from IARPA and ODNI for their followup paper Labeled Faces in the Wild: Updates and New Reporting Procedures via IARPA contract number 2014-14071600010
          • The dataset includes 2 images of George Tenet, the former Director of Central Intelligence (DCI) for the Central Intelligence Agency whose facial biometrics were eventually used to help train facial recognition software in China and Russia
            • @@ -94,7 +94,10 @@ imageio.imwrite('lfw_montage_960.jpg', montage)

          Supplementary Material

          Text and graphics ©Adam Harvey / megapixels.cc

          -
          +
          -- cgit v1.2.3-70-g09d2 From 0f8db17624b4b85f3fab72b4c1037d286fe047c8 Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Wed, 27 Feb 2019 23:56:07 +0100 Subject: footnotes working --- site/assets/css/css.css | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'site') diff --git a/site/assets/css/css.css b/site/assets/css/css.css index 8b4241ea..ebe5193e 100644 --- a/site/assets/css/css.css +++ b/site/assets/css/css.css @@ -211,7 +211,7 @@ p { line-height: 2; } .content a { - color: #ddd; + color: #ff0; transition: color 0.2s cubic-bezier(0,0,1,1); } .content a:hover { -- cgit v1.2.3-70-g09d2 From 6ad53c878f22e44836f8eb1933882ad61411da89 Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Wed, 27 Feb 2019 23:57:13 +0100 Subject: yellow links --- site/assets/css/css.css | 1 + 1 file changed, 1 insertion(+) (limited to 'site') diff --git a/site/assets/css/css.css b/site/assets/css/css.css index ebe5193e..0afa3725 100644 --- a/site/assets/css/css.css +++ b/site/assets/css/css.css @@ -110,6 +110,7 @@ footer { color: #888; font-size: 9pt; padding: 20px 75px 20px; + font-family: "Roboto", sans-serif; } footer > div { display: flex; -- cgit v1.2.3-70-g09d2 From be343b455d3f65f268602efe0a93302a1748f493 Mon Sep 17 00:00:00 2001 From: Adam Harvey Date: Thu, 28 Feb 2019 00:13:32 +0100 Subject: add lfw bg --- .../50_people_one_question/assets/background.gif | Bin 0 -> 41564 bytes site/content/pages/datasets/lfw/assets/background.jpg | Bin 0 -> 239321 bytes site/content/pages/datasets/mars/assets/background.jpg | Bin 0 -> 216396 bytes .../content/pages/datasets/viper/assets/background.jpg | Bin 0 -> 203679 bytes site/content/pages/research/00_introduction/index.md | 9 ++++++--- 5 files changed, 6 insertions(+), 3 deletions(-) create mode 100644 site/content/pages/datasets/50_people_one_question/assets/background.gif create mode 100644 site/content/pages/datasets/lfw/assets/background.jpg create mode 100644 site/content/pages/datasets/mars/assets/background.jpg create mode 100644 site/content/pages/datasets/viper/assets/background.jpg (limited to 'site') diff --git a/site/content/pages/datasets/50_people_one_question/assets/background.gif b/site/content/pages/datasets/50_people_one_question/assets/background.gif new file mode 100644 index 00000000..a0539bbb Binary files /dev/null and b/site/content/pages/datasets/50_people_one_question/assets/background.gif differ diff --git a/site/content/pages/datasets/lfw/assets/background.jpg b/site/content/pages/datasets/lfw/assets/background.jpg new file mode 100644 index 00000000..64d61c35 Binary files /dev/null and b/site/content/pages/datasets/lfw/assets/background.jpg differ diff --git a/site/content/pages/datasets/mars/assets/background.jpg b/site/content/pages/datasets/mars/assets/background.jpg new file mode 100644 index 00000000..9c16c26d Binary files /dev/null and b/site/content/pages/datasets/mars/assets/background.jpg differ diff --git a/site/content/pages/datasets/viper/assets/background.jpg b/site/content/pages/datasets/viper/assets/background.jpg new file mode 100644 index 00000000..db0b2857 Binary files /dev/null and b/site/content/pages/datasets/viper/assets/background.jpg differ diff --git a/site/content/pages/research/00_introduction/index.md b/site/content/pages/research/00_introduction/index.md index 1b784768..6fec7ab5 100644 --- a/site/content/pages/research/00_introduction/index.md +++ b/site/content/pages/research/00_introduction/index.md @@ -16,14 +16,17 @@ authors: Megapixels + Author: Adam Harvey ------ +Ever since the first computational facial recognition research project by the CIA in the early 1960s, data has always played a vital role in the development of our biometric future. Without facial recognition datasets there would be no facial recognition. Datasets are an indispensable part of any artificial intelligence system because, as Geoffrey Hinton points out: +> Our relationship to computers has changed. Instead of programming them, we now show them and they figure it out. - [Geoffrey Hinton](https://www.youtube.com/watch?v=-eyhCTvrEtE) -Ignore content below these lines +Algorithms learn from datasets. And we program algorithms by building datasets. But datasets aren't like code. There's no programming language made of data except for the data itself. ----- -Ever since the first computational facial recognition research project by the CIA in the early 1960s, data has always played a vital role in the development of our biometric future. Without facial recognition datasets there would be no facial recognition. Datasets are an indispensable part of any artificial intelligence system because, as Geoffrey Hinton points out, "we no longer program computers with code, we program them with data". +Ignore content below these lines +----- + It was the early 2000s. Face recognition was new and no one seemed sure exactly how well it was going to perform in practice. In theory, face recognition was poised to be a game changer, a force multiplier, a strategic military advantage, a way to make cities safer and to secure borders. This was the future John Ashcroft demanded with the Total Information Awareness act of the 2003 and that spooks had dreamed of for decades. It was a future that academics at Carnegie Mellon Universtiy and Colorado State University would help build. It was also a future that celebrities would play a significant role in building. And to the surprise of ordinary Internet users like myself and perhaps you, it was a future that millions of Internet users would unwittingly play role in creating. -- cgit v1.2.3-70-g09d2 From 85852a4cf2eb5cc364afd85d809cca32e998175d Mon Sep 17 00:00:00 2001 From: Adam Harvey Date: Thu, 28 Feb 2019 14:29:15 +0100 Subject: add images --- .../50_people_one_question/assets/index.jpg | Bin 0 -> 2981 bytes .../50_people_one_question/assets/index_02.jpg | Bin 0 -> 3064 bytes .../pages/datasets/brainwash/assets/index.jpg | Bin 0 -> 3056 bytes .../pages/datasets/duke_mtmc/assets/index.jpg | Bin 0 -> 2436 bytes site/content/pages/datasets/facebook/index.md | 32 +++++ site/content/pages/datasets/helen/assets/index.jpg | Bin 0 -> 3243 bytes .../datasets/hrt_transgender/assets/index.jpg | Bin 0 -> 3085 bytes site/content/pages/datasets/lfw/assets/index.jpg | Bin 0 -> 25306 bytes .../pages/datasets/lfw/assets/lfw_index.gif | Bin 148777 -> 0 bytes .../pages/datasets/lfw/assets/lfw_montage.jpg | Bin 358848 -> 0 bytes .../pages/datasets/lfw/assets/lfw_synthetic.jpg | Bin 159745 -> 0 bytes site/content/pages/datasets/mars/assets/index.jpg | Bin 0 -> 23722 bytes .../content/pages/datasets/pubfig/assets/index.jpg | Bin 0 -> 20533 bytes site/content/pages/datasets/ytmu/assets/index.jpg | Bin 0 -> 6489 bytes .../pages/datasets/ytmu/assets/index_02.jpg | Bin 0 -> 5684 bytes .../pages/datasets/ytmu/assets/index_03.jpg | Bin 0 -> 5974 bytes .../research/01_from_1_to_100_pixels/index.md | 4 +- site/datasets/final/ijb_c_sample.csv | 141 +++++++++++++++++++++ 18 files changed, 176 insertions(+), 1 deletion(-) create mode 100644 site/content/pages/datasets/50_people_one_question/assets/index.jpg create mode 100644 site/content/pages/datasets/50_people_one_question/assets/index_02.jpg create mode 100644 site/content/pages/datasets/brainwash/assets/index.jpg create mode 100644 site/content/pages/datasets/duke_mtmc/assets/index.jpg create mode 100644 site/content/pages/datasets/facebook/index.md create mode 100644 site/content/pages/datasets/helen/assets/index.jpg create mode 100644 site/content/pages/datasets/hrt_transgender/assets/index.jpg create mode 100644 site/content/pages/datasets/lfw/assets/index.jpg delete mode 100644 site/content/pages/datasets/lfw/assets/lfw_index.gif delete mode 100644 site/content/pages/datasets/lfw/assets/lfw_montage.jpg delete mode 100644 site/content/pages/datasets/lfw/assets/lfw_synthetic.jpg create mode 100644 site/content/pages/datasets/mars/assets/index.jpg create mode 100644 site/content/pages/datasets/pubfig/assets/index.jpg create mode 100644 site/content/pages/datasets/ytmu/assets/index.jpg create mode 100644 site/content/pages/datasets/ytmu/assets/index_02.jpg create mode 100644 site/content/pages/datasets/ytmu/assets/index_03.jpg create mode 100644 site/datasets/final/ijb_c_sample.csv (limited to 'site') diff --git a/site/content/pages/datasets/50_people_one_question/assets/index.jpg b/site/content/pages/datasets/50_people_one_question/assets/index.jpg new file mode 100644 index 00000000..a79c7739 Binary files /dev/null and b/site/content/pages/datasets/50_people_one_question/assets/index.jpg differ diff --git a/site/content/pages/datasets/50_people_one_question/assets/index_02.jpg b/site/content/pages/datasets/50_people_one_question/assets/index_02.jpg new file mode 100644 index 00000000..c331ea5c Binary files /dev/null and b/site/content/pages/datasets/50_people_one_question/assets/index_02.jpg differ diff --git a/site/content/pages/datasets/brainwash/assets/index.jpg b/site/content/pages/datasets/brainwash/assets/index.jpg new file mode 100644 index 00000000..7d6230e1 Binary files /dev/null and b/site/content/pages/datasets/brainwash/assets/index.jpg differ diff --git a/site/content/pages/datasets/duke_mtmc/assets/index.jpg b/site/content/pages/datasets/duke_mtmc/assets/index.jpg new file mode 100644 index 00000000..6651c15c Binary files /dev/null and b/site/content/pages/datasets/duke_mtmc/assets/index.jpg differ diff --git a/site/content/pages/datasets/facebook/index.md b/site/content/pages/datasets/facebook/index.md new file mode 100644 index 00000000..6e3857fd --- /dev/null +++ b/site/content/pages/datasets/facebook/index.md @@ -0,0 +1,32 @@ +------------ + +status: published +title: Facebook +desc: TBD +subdesc: TBD +image: assets/background.jpg +caption: TBD +slug: facebook +published: 2019-2-23 +updated: 2019-2-23 +color: #aaaaff +authors: Adam Harvey + +------------ + +### Statistics + ++ Years: 2002-2004 ++ Images: 13,233 ++ Identities: 5,749 ++ Origin: Yahoo News Images ++ Funding: (Possibly, partially CIA) + +---- + +Ignore content below these lines + +--- + + +- Tool to create face datasets from Facebook diff --git a/site/content/pages/datasets/helen/assets/index.jpg b/site/content/pages/datasets/helen/assets/index.jpg new file mode 100644 index 00000000..37cb5882 Binary files /dev/null and b/site/content/pages/datasets/helen/assets/index.jpg differ diff --git a/site/content/pages/datasets/hrt_transgender/assets/index.jpg b/site/content/pages/datasets/hrt_transgender/assets/index.jpg new file mode 100644 index 00000000..4f3307c4 Binary files /dev/null and b/site/content/pages/datasets/hrt_transgender/assets/index.jpg differ diff --git a/site/content/pages/datasets/lfw/assets/index.jpg b/site/content/pages/datasets/lfw/assets/index.jpg new file mode 100644 index 00000000..1df0c4f6 Binary files /dev/null and b/site/content/pages/datasets/lfw/assets/index.jpg differ diff --git a/site/content/pages/datasets/lfw/assets/lfw_index.gif b/site/content/pages/datasets/lfw/assets/lfw_index.gif deleted file mode 100644 index 63a5c423..00000000 Binary files a/site/content/pages/datasets/lfw/assets/lfw_index.gif and /dev/null differ diff --git a/site/content/pages/datasets/lfw/assets/lfw_montage.jpg b/site/content/pages/datasets/lfw/assets/lfw_montage.jpg deleted file mode 100644 index 8d73c417..00000000 Binary files a/site/content/pages/datasets/lfw/assets/lfw_montage.jpg and /dev/null differ diff --git a/site/content/pages/datasets/lfw/assets/lfw_synthetic.jpg b/site/content/pages/datasets/lfw/assets/lfw_synthetic.jpg deleted file mode 100644 index c2a34043..00000000 Binary files a/site/content/pages/datasets/lfw/assets/lfw_synthetic.jpg and /dev/null differ diff --git a/site/content/pages/datasets/mars/assets/index.jpg b/site/content/pages/datasets/mars/assets/index.jpg new file mode 100644 index 00000000..9e527fbd Binary files /dev/null and b/site/content/pages/datasets/mars/assets/index.jpg differ diff --git a/site/content/pages/datasets/pubfig/assets/index.jpg b/site/content/pages/datasets/pubfig/assets/index.jpg new file mode 100644 index 00000000..be14f27c Binary files /dev/null and b/site/content/pages/datasets/pubfig/assets/index.jpg differ diff --git a/site/content/pages/datasets/ytmu/assets/index.jpg b/site/content/pages/datasets/ytmu/assets/index.jpg new file mode 100644 index 00000000..6df15db5 Binary files /dev/null and b/site/content/pages/datasets/ytmu/assets/index.jpg differ diff --git a/site/content/pages/datasets/ytmu/assets/index_02.jpg b/site/content/pages/datasets/ytmu/assets/index_02.jpg new file mode 100644 index 00000000..30c863f6 Binary files /dev/null and b/site/content/pages/datasets/ytmu/assets/index_02.jpg differ diff --git a/site/content/pages/datasets/ytmu/assets/index_03.jpg b/site/content/pages/datasets/ytmu/assets/index_03.jpg new file mode 100644 index 00000000..20ccae90 Binary files /dev/null and b/site/content/pages/datasets/ytmu/assets/index_03.jpg differ diff --git a/site/content/pages/research/01_from_1_to_100_pixels/index.md b/site/content/pages/research/01_from_1_to_100_pixels/index.md index 0123fffe..29204168 100644 --- a/site/content/pages/research/01_from_1_to_100_pixels/index.md +++ b/site/content/pages/research/01_from_1_to_100_pixels/index.md @@ -45,4 +45,6 @@ Find specific cases of facial resolution being used in legal cases, forensic inv - NIST report on sres states several resolutions - "Results show that the tested face recognition systems yielded similar performance for query sets with eye-to-eye distance from 60 pixels to 30 pixels" [^nist_sres] -[^nist_sres]: NIST 906932. Performance Assessment of Face Recognition Using Super-Resolution. Shuowen Hu, Robert Maschal, S. Susan Young, Tsai Hong Hong, Jonathon P. Phillips \ No newline at end of file +[^nist_sres]: NIST 906932. Performance Assessment of Face Recognition Using Super-Resolution. Shuowen Hu, Robert Maschal, S. Susan Young, Tsai Hong Hong, Jonathon P. Phillips + +- "Note that we only keep the images with a minimal side length of 80 pixels." and "a face will be labeled as “Ignore” if it is very difficult to be detected due to blurring, severe deformation and unrecognizable eyes, or the side length of its bounding box is less than 32 pixels." Ge_Detecting_Masked_Faces_CVPR_2017_paper.pdf \ No newline at end of file diff --git a/site/datasets/final/ijb_c_sample.csv b/site/datasets/final/ijb_c_sample.csv new file mode 100644 index 00000000..15bfccab --- /dev/null +++ b/site/datasets/final/ijb_c_sample.csv @@ -0,0 +1,141 @@ +index,dataset_name,key,lat,lng,loc,loc_type,paper_id,paper_type,paper_url,title,year +0,IJB-A,ijb_c,0.0,0.0,,,140c95e53c619eac594d70f6369f518adfea12ef,main,http://www.cv-foundation.org/openaccess/content_cvpr_2015/app/1B_089_ext.pdf,Pushing the frontiers of unconstrained face detection and recognition: IARPA Janus Benchmark A,2015 +1,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,872dfdeccf99bbbed7c8f1ea08afb2d713ebe085,citation,https://arxiv.org/pdf/1703.09507.pdf,L2-constrained Softmax Loss for Discriminative Face Verification,2017 +2,IJB-A,ijb_c,38.8920756,-104.79716389,"University of Colorado, Colorado Springs",edu,146a7ecc7e34b85276dd0275c337eff6ba6ef8c0,citation,https://arxiv.org/pdf/1611.06158v1.pdf,AFFACT: Alignment-free facial attribute classification technique,2017 +3,IJB-A,ijb_c,51.7534538,-1.25400997,University of Oxford,edu,313d5eba97fe064bdc1f00b7587a4b3543ef712a,citation,https://pdfs.semanticscholar.org/cb7f/93467b0ec1afd43d995e511f5d7bf052a5af.pdf,Compact Deep Aggregation for Set Retrieval,2018 +4,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,5865b6d83ba6dbbf9167f1481e9339c2ef1d1f6b,citation,https://doi.org/10.1109/ICPR.2016.7900278,Regularized metric adaptation for unconstrained face verification,2016 +5,IJB-A,ijb_c,37.4102193,-122.05965487,Carnegie Mellon University,edu,48a9241edda07252c1aadca09875fabcfee32871,citation,https://arxiv.org/pdf/1611.08657v5.pdf,Convolutional Experts Constrained Local Model for Facial Landmark Detection,2017 +6,IJB-A,ijb_c,42.718568,-84.47791571,Michigan State University,edu,86204fc037936754813b91898377e8831396551a,citation,https://arxiv.org/pdf/1709.01442.pdf,Dense Face Alignment,2017 +7,IJB-A,ijb_c,22.57423855,88.4337303,"Institute of Engineering and Management, Kolkata, India",edu,b2cb335ded99b10f37002d09753bd5a6ea522ef1,citation,https://doi.org/10.1109/ISBA.2017.7947679,Analysis of adaptability of deep features for verifying blurred and cross-resolution images,2017 +8,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,b2cb335ded99b10f37002d09753bd5a6ea522ef1,citation,https://doi.org/10.1109/ISBA.2017.7947679,Analysis of adaptability of deep features for verifying blurred and cross-resolution images,2017 +9,IJB-A,ijb_c,45.7835966,4.7678948,École Centrale de Lyon,edu,486840f4f524e97f692a7f6b42cd19019ee71533,citation,https://arxiv.org/pdf/1703.08388v2.pdf,DeepVisage: Making Face Recognition Simple Yet With Powerful Generalization Skills,2017 +10,IJB-A,ijb_c,48.832493,2.267474,Safran Identity and Security,company,486840f4f524e97f692a7f6b42cd19019ee71533,citation,https://arxiv.org/pdf/1703.08388v2.pdf,DeepVisage: Making Face Recognition Simple Yet With Powerful Generalization Skills,2017 +11,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,2d748f8ee023a5b1fbd50294d176981ded4ad4ee,citation,http://pdfs.semanticscholar.org/2d74/8f8ee023a5b1fbd50294d176981ded4ad4ee.pdf,Triplet Similarity Embedding for Face Verification,2016 +12,IJB-A,ijb_c,38.99203005,-76.9461029,University of Maryland College Park,edu,f7824758800a7b1a386db5bd35f84c81454d017a,citation,https://arxiv.org/pdf/1702.05085.pdf,KEPLER: Keypoint and Pose Estimation of Unconstrained Faces by Learning Efficient H-CNN Regressors,2017 +13,IJB-A,ijb_c,42.718568,-84.47791571,Michigan State University,edu,02467703b6e087799e04e321bea3a4c354c5487d,citation,http://doi.ieeecomputersociety.org/10.1109/CVPRW.2016.27,Grouper: Optimizing Crowdsourced Face Annotations,2016 +14,IJB-A,ijb_c,39.329053,-76.619425,Johns Hopkins University,edu,377f2b65e6a9300448bdccf678cde59449ecd337,citation,https://arxiv.org/pdf/1804.10275.pdf,Pushing the Limits of Unconstrained Face Detection: a Challenge Dataset and Baseline Results,2018 +15,IJB-A,ijb_c,40.47913175,-74.43168868,Rutgers University,edu,377f2b65e6a9300448bdccf678cde59449ecd337,citation,https://arxiv.org/pdf/1804.10275.pdf,Pushing the Limits of Unconstrained Face Detection: a Challenge Dataset and Baseline Results,2018 +16,IJB-A,ijb_c,42.718568,-84.47791571,Michigan State University,edu,cd55fb30737625e86454a2861302b96833ed549d,citation,http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7139094,Annotating Unconstrained Face Imagery: A scalable approach,2015 +17,IJB-A,ijb_c,38.95187,-77.363259,"Noblis, Falls Church, VA, U.S.A.",company,cd55fb30737625e86454a2861302b96833ed549d,citation,http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7139094,Annotating Unconstrained Face Imagery: A scalable approach,2015 +18,IJB-A,ijb_c,46.0501558,14.46907327,University of Ljubljana,edu,5226296884b3e151ce317a37f94827dbda0b9d16,citation,https://doi.org/10.1109/IWBF.2016.7449690,Deep pair-wise similarity learning for face recognition,2016 +19,IJB-A,ijb_c,39.9601488,116.35193921,Beijing University of Posts and Telecommunications,edu,80be8624771104ff4838dcba9629bacfe6b3ea09,citation,http://www.ifp.illinois.edu/~moulin/Papers/ECCV14-jiwen.pdf,Simultaneous Feature and Dictionary Learning for Image Set Based Face Recognition,2014 +20,IJB-A,ijb_c,1.3484104,103.68297965,Nanyang Technological University,edu,80be8624771104ff4838dcba9629bacfe6b3ea09,citation,http://www.ifp.illinois.edu/~moulin/Papers/ECCV14-jiwen.pdf,Simultaneous Feature and Dictionary Learning for Image Set Based Face Recognition,2014 +21,IJB-A,ijb_c,40.11116745,-88.22587665,"University of Illinois, Urbana-Champaign",edu,80be8624771104ff4838dcba9629bacfe6b3ea09,citation,http://www.ifp.illinois.edu/~moulin/Papers/ECCV14-jiwen.pdf,Simultaneous Feature and Dictionary Learning for Image Set Based Face Recognition,2014 +22,IJB-A,ijb_c,22.304572,114.17976285,Hong Kong Polytechnic University,edu,50b58becaf67e92a6d9633e0eea7d352157377c3,citation,https://pdfs.semanticscholar.org/50b5/8becaf67e92a6d9633e0eea7d352157377c3.pdf,Dependency-Aware Attention Control for Unconstrained Face Recognition with Image Sets,2018 +23,IJB-A,ijb_c,39.9601488,116.35193921,Beijing University of Posts and Telecommunications,edu,cd6aaa37fffd0b5c2320f386be322b8adaa1cc68,citation,https://arxiv.org/pdf/1804.06655.pdf,Deep Face Recognition: A Survey,2018 +24,IJB-A,ijb_c,39.9082804,116.2458527,University of Chinese Academy of Sciences,edu,ac2881bdf7b57dc1672a17b221d68a438d79fce8,citation,https://arxiv.org/pdf/1806.08472.pdf,Learning a High Fidelity Pose Invariant Model for High-resolution Face Frontalization,2018 +25,IJB-A,ijb_c,40.0044795,116.370238,Chinese Academy of Sciences,edu,72a7eb68f0955564e1ceafa75aeeb6b5bbb14e7e,citation,https://pdfs.semanticscholar.org/72a7/eb68f0955564e1ceafa75aeeb6b5bbb14e7e.pdf,Face Recognition with Contrastive Convolution,2018 +26,IJB-A,ijb_c,39.9082804,116.2458527,University of Chinese Academy of Sciences,edu,72a7eb68f0955564e1ceafa75aeeb6b5bbb14e7e,citation,https://pdfs.semanticscholar.org/72a7/eb68f0955564e1ceafa75aeeb6b5bbb14e7e.pdf,Face Recognition with Contrastive Convolution,2018 +27,IJB-A,ijb_c,42.3889785,-72.5286987,University of Massachusetts,edu,368e99f669ea5fd395b3193cd75b301a76150f9d,citation,https://arxiv.org/pdf/1506.01342.pdf,One-to-many face recognition with bilinear CNNs,2016 +28,IJB-A,ijb_c,32.77824165,34.99565673,Open University of Israel,edu,1e6ed6ca8209340573a5e907a6e2e546a3bf2d28,citation,http://arxiv.org/pdf/1607.01450v1.pdf,Pooling Faces: Template Based Face Recognition with Pooled Face Images,2016 +29,IJB-A,ijb_c,38.88140235,121.52281098,Dalian University of Technology,edu,052f994898c79529955917f3dfc5181586282cf8,citation,https://arxiv.org/pdf/1708.02191.pdf,Unsupervised Domain Adaptation for Face Recognition in Unlabeled Videos,2017 +30,IJB-A,ijb_c,32.9820799,-96.7566278,University of Texas at Dallas,edu,4e8168fbaa615009d1618a9d6552bfad809309e9,citation,http://pdfs.semanticscholar.org/4e81/68fbaa615009d1618a9d6552bfad809309e9.pdf,Deep Convolutional Neural Network Features and the Original Image,2016 +31,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,4e8168fbaa615009d1618a9d6552bfad809309e9,citation,http://pdfs.semanticscholar.org/4e81/68fbaa615009d1618a9d6552bfad809309e9.pdf,Deep Convolutional Neural Network Features and the Original Image,2016 +32,IJB-A,ijb_c,29.7207902,-95.34406271,University of Houston,edu,3cb2841302af1fb9656f144abc79d4f3d0b27380,citation,https://pdfs.semanticscholar.org/3cb2/841302af1fb9656f144abc79d4f3d0b27380.pdf,When 3 D-Aided 2 D Face Recognition Meets Deep Learning : An extended UR 2 D for Pose-Invariant Face Recognition,2017 +33,IJB-A,ijb_c,24.4469025,54.3942563,Khalifa University,edu,0c1d85a197a1f5b7376652a485523e616a406273,citation,http://doi.ieeecomputersociety.org/10.1109/CVPR.2017.169,Joint Registration and Representation Learning for Unconstrained Face Identification,2017 +34,IJB-A,ijb_c,-35.23656905,149.08446994,University of Canberra,edu,0c1d85a197a1f5b7376652a485523e616a406273,citation,http://doi.ieeecomputersociety.org/10.1109/CVPR.2017.169,Joint Registration and Representation Learning for Unconstrained Face Identification,2017 +35,IJB-A,ijb_c,32.77824165,34.99565673,Open University of Israel,edu,c75e6ce54caf17b2780b4b53f8d29086b391e839,citation,https://arxiv.org/pdf/1802.00542.pdf,"ExpNet: Landmark-Free, Deep, 3D Facial Expressions",2018 +36,IJB-A,ijb_c,42.718568,-84.47791571,Michigan State University,edu,450c6a57f19f5aa45626bb08d7d5d6acdb863b4b,citation,https://arxiv.org/pdf/1805.00611.pdf,Towards Interpretable Face Recognition,2018 +37,IJB-A,ijb_c,51.7534538,-1.25400997,University of Oxford,edu,30180f66d5b4b7c0367e4b43e2b55367b72d6d2a,citation,http://www.robots.ox.ac.uk/~vgg/publications/2017/Crosswhite17/crosswhite17.pdf,Template Adaptation for Face Verification and Identification,2017 +38,IJB-A,ijb_c,29.7207902,-95.34406271,University of Houston,edu,8334da483f1986aea87b62028672836cb3dc6205,citation,https://arxiv.org/pdf/1805.06306.pdf,Fully Associative Patch-Based 1-to-N Matcher for Face Recognition,2018 +39,IJB-A,ijb_c,-33.8809651,151.20107299,University of Technology Sydney,edu,3b64efa817fd609d525c7244a0e00f98feacc8b4,citation,http://doi.acm.org/10.1145/2845089,A Comprehensive Survey on Pose-Invariant Face Recognition,2016 +40,IJB-A,ijb_c,40.9153196,-73.1270626,Stony Brook University,edu,6fbb179a4ad39790f4558dd32316b9f2818cd106,citation,http://pdfs.semanticscholar.org/6fbb/179a4ad39790f4558dd32316b9f2818cd106.pdf,Input Aggregated Network for Face Video Representation,2016 +41,IJB-A,ijb_c,38.8920756,-104.79716389,"University of Colorado, Colorado Springs",edu,d4f1eb008eb80595bcfdac368e23ae9754e1e745,citation,https://arxiv.org/pdf/1708.02337.pdf,Unconstrained Face Detection and Open-Set Face Recognition Challenge,2017 +42,IJB-A,ijb_c,33.5866784,-101.87539204,Electrical and Computer Engineering,edu,ebb3d5c70bedf2287f9b26ac0031004f8f617b97,citation,https://doi.org/10.1109/MSP.2017.2764116,"Deep Learning for Understanding Faces: Machines May Be Just as Good, or Better, than Humans",2018 +43,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,ebb3d5c70bedf2287f9b26ac0031004f8f617b97,citation,https://doi.org/10.1109/MSP.2017.2764116,"Deep Learning for Understanding Faces: Machines May Be Just as Good, or Better, than Humans",2018 +44,IJB-A,ijb_c,34.0224149,-118.28634407,University of Southern California,edu,d28d32af7ef9889ef9cb877345a90ea85e70f7f1,citation,http://doi.ieeecomputersociety.org/10.1109/FG.2017.84,Local-Global Landmark Confidences for Face Recognition,2017 +45,IJB-A,ijb_c,37.4102193,-122.05965487,Carnegie Mellon University,edu,d28d32af7ef9889ef9cb877345a90ea85e70f7f1,citation,http://doi.ieeecomputersociety.org/10.1109/FG.2017.84,Local-Global Landmark Confidences for Face Recognition,2017 +46,IJB-A,ijb_c,51.5247272,-0.03931035,Queen Mary University of London,edu,a29566375836f37173ccaffa47dea25eb1240187,citation,https://arxiv.org/pdf/1809.09409.pdf,Vehicle Re-Identification in Context,2018 +47,IJB-A,ijb_c,34.0224149,-118.28634407,University of Southern California,edu,29f298dd5f806c99951cb434834bc8dcc765df18,citation,https://doi.org/10.1109/ICPR.2016.7899837,Computationally efficient template-based face recognition,2016 +48,IJB-A,ijb_c,51.49887085,-0.17560797,Imperial College London,edu,54bb25a213944b08298e4e2de54f2ddea890954a,citation,http://openaccess.thecvf.com/content_cvpr_2017_workshops/w33/papers/Moschoglou_AgeDB_The_First_CVPR_2017_paper.pdf,"AgeDB: The First Manually Collected, In-the-Wild Age Database",2017 +49,IJB-A,ijb_c,51.59029705,-0.22963221,Middlesex University,edu,54bb25a213944b08298e4e2de54f2ddea890954a,citation,http://openaccess.thecvf.com/content_cvpr_2017_workshops/w33/papers/Moschoglou_AgeDB_The_First_CVPR_2017_paper.pdf,"AgeDB: The First Manually Collected, In-the-Wild Age Database",2017 +50,IJB-A,ijb_c,50.8142701,8.771435,Philipps-Universität Marburg,edu,5981c309bd0ffd849c51b1d8a2ccc481a8ec2f5c,citation,https://doi.org/10.1109/ICT.2017.7998256,SmartFace: Efficient face detection on smartphones for wireless on-demand emergency networks,2017 +51,IJB-A,ijb_c,42.718568,-84.47791571,Michigan State University,edu,a2b4a6c6b32900a066d0257ae6d4526db872afe2,citation,http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8272466,Learning Face Image Quality From Human Assessments,2018 +52,IJB-A,ijb_c,39.9601488,116.35193921,Beijing University of Posts and Telecommunications,edu,3dfb822e16328e0f98a47209d7ecd242e4211f82,citation,https://arxiv.org/pdf/1708.08197.pdf,Cross-Age LFW: A Database for Studying Cross-Age Face Recognition in Unconstrained Environments,2017 +53,IJB-A,ijb_c,47.6423318,-122.1369302,Microsoft,company,291265db88023e92bb8c8e6390438e5da148e8f5,citation,http://pdfs.semanticscholar.org/4603/cb8e05258bb0572ae912ad20903b8f99f4b1.pdf,MS-Celeb-1M: A Dataset and Benchmark for Large-Scale Face Recognition,2016 +54,IJB-A,ijb_c,42.718568,-84.47791571,Michigan State University,edu,d29eec5e047560627c16803029d2eb8a4e61da75,citation,http://pdfs.semanticscholar.org/d29e/ec5e047560627c16803029d2eb8a4e61da75.pdf,Feature Transfer Learning for Deep Face Recognition with Long-Tail Data,2018 +55,IJB-A,ijb_c,36.20304395,117.05842113,Tianjin University,edu,5180df9d5eb26283fb737f491623395304d57497,citation,https://arxiv.org/pdf/1804.10899.pdf,Scalable Angular Discriminative Deep Metric Learning for Face Recognition,2018 +56,IJB-A,ijb_c,22.42031295,114.20788644,Chinese University of Hong Kong,edu,abdd17e411a7bfe043f280abd4e560a04ab6e992,citation,https://arxiv.org/pdf/1803.00839.pdf,Pose-Robust Face Recognition via Deep Residual Equivariant Mapping,2018 +57,IJB-A,ijb_c,28.5456282,77.2731505,"IIIT Delhi, India",edu,3cf1f89d73ca4b25399c237ed3e664a55cd273a2,citation,https://arxiv.org/pdf/1710.02914.pdf,Face Sketch Matching via Coupled Deep Transform Learning,2017 +58,IJB-A,ijb_c,-27.49741805,153.01316956,University of Queensland,edu,f27fd2a1bc229c773238f1912db94991b8bf389a,citation,https://doi.org/10.1109/IVCNZ.2016.7804414,How do you develop a face detector for the unconstrained environment?,2016 +59,IJB-A,ijb_c,39.86742125,32.73519072,Hacettepe University,edu,9865fe20df8fe11717d92b5ea63469f59cf1635a,citation,https://arxiv.org/pdf/1805.07566.pdf,Wildest Faces: Face Detection and Recognition in Violent Settings,2018 +60,IJB-A,ijb_c,39.87549675,32.78553506,Middle East Technical University,edu,9865fe20df8fe11717d92b5ea63469f59cf1635a,citation,https://arxiv.org/pdf/1805.07566.pdf,Wildest Faces: Face Detection and Recognition in Violent Settings,2018 +61,IJB-A,ijb_c,28.2290209,112.99483204,"National University of Defense Technology, China",edu,c1cc2a2a1ab66f6c9c6fabe28be45d1440a57c3d,citation,https://pdfs.semanticscholar.org/aae7/a5182e59f44b7bb49f61999181ce011f800b.pdf,Dual-Agent GANs for Photorealistic and Identity Preserving Profile Face Synthesis,2017 +62,IJB-A,ijb_c,1.2962018,103.77689944,National University of Singapore,edu,c1cc2a2a1ab66f6c9c6fabe28be45d1440a57c3d,citation,https://pdfs.semanticscholar.org/aae7/a5182e59f44b7bb49f61999181ce011f800b.pdf,Dual-Agent GANs for Photorealistic and Identity Preserving Profile Face Synthesis,2017 +63,IJB-A,ijb_c,17.4454957,78.34854698,International Institute of Information Technology,edu,f5eb411217f729ad7ae84bfd4aeb3dedb850206a,citation,https://pdfs.semanticscholar.org/f5eb/411217f729ad7ae84bfd4aeb3dedb850206a.pdf,Tackling Low Resolution for Better Scene Understanding,2018 +64,IJB-A,ijb_c,40.51865195,-74.44099801,State University of New Jersey,edu,96e731e82b817c95d4ce48b9e6b08d2394937cf8,citation,http://arxiv.org/pdf/1508.01722v2.pdf,Unconstrained face verification using deep CNN features,2016 +65,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,96e731e82b817c95d4ce48b9e6b08d2394937cf8,citation,http://arxiv.org/pdf/1508.01722v2.pdf,Unconstrained face verification using deep CNN features,2016 +66,IJB-A,ijb_c,32.77824165,34.99565673,Open University of Israel,edu,870433ba89d8cab1656e57ac78f1c26f4998edfb,citation,http://doi.ieeecomputersociety.org/10.1109/CVPR.2017.163,Regressing Robust and Discriminative 3D Morphable Models with a Very Deep Neural Network,2017 +67,IJB-A,ijb_c,55.6801502,12.572327,University of Copenhagen,edu,3dfd94d3fad7e17f52a8ae815eb9cc5471172bc0,citation,http://pdfs.semanticscholar.org/3dfd/94d3fad7e17f52a8ae815eb9cc5471172bc0.pdf,Face2Text: Collecting an Annotated Image Description Corpus for the Generation of Rich Face Descriptions,2018 +68,IJB-A,ijb_c,35.9023226,14.4834189,University of Malta,edu,3dfd94d3fad7e17f52a8ae815eb9cc5471172bc0,citation,http://pdfs.semanticscholar.org/3dfd/94d3fad7e17f52a8ae815eb9cc5471172bc0.pdf,Face2Text: Collecting an Annotated Image Description Corpus for the Generation of Rich Face Descriptions,2018 +69,IJB-A,ijb_c,34.0224149,-118.28634407,University of Southern California,edu,6341274aca0c2977c3e1575378f4f2126aa9b050,citation,http://arxiv.org/pdf/1609.03536v1.pdf,A multi-scale cascade fully convolutional network face detector,2016 +70,IJB-A,ijb_c,41.70456775,-86.23822026,University of Notre Dame,edu,17479e015a2dcf15d40190e06419a135b66da4e0,citation,https://arxiv.org/pdf/1610.08119.pdf,Predicting First Impressions With Deep Learning,2017 +71,IJB-A,ijb_c,37.4102193,-122.05965487,Carnegie Mellon University,edu,a0b1990dd2b4cd87e4fd60912cc1552c34792770,citation,https://pdfs.semanticscholar.org/a0b1/990dd2b4cd87e4fd60912cc1552c34792770.pdf,Deep Constrained Local Models for Facial Landmark Detection,2016 +72,IJB-A,ijb_c,30.642769,104.06751175,"Sichuan University, Chengdu",edu,772474b5b0c90629f4d9c223fd9c1ef45e1b1e66,citation,https://doi.org/10.1109/BTAS.2017.8272716,Multi-dim: A multi-dimensional face database towards the application of 3D technology in real-world scenarios,2017 +73,IJB-A,ijb_c,38.8920756,-104.79716389,"University of Colorado, Colorado Springs",edu,4b3f425274b0c2297d136f8833a31866db2f2aec,citation,http://doi.ieeecomputersociety.org/10.1109/CVPRW.2017.85,Toward Open-Set Face Recognition,2017 +74,IJB-A,ijb_c,56.46255985,84.95565495,Tomsk Polytechnic University,edu,17ded725602b4329b1c494bfa41527482bf83a6f,citation,http://pdfs.semanticscholar.org/cb10/434a5d68ffbe9ed0498771192564ecae8894.pdf,Compact Convolutional Neural Network Cascade for Face Detection,2015 +75,IJB-A,ijb_c,37.3351908,-121.88126008,San Jose State University,edu,14b016c7a87d142f4b9a0e6dc470dcfc073af517,citation,http://ws680.nist.gov/publication/get_pdf.cfm?pub_id=918912,Modest proposals for improving biometric recognition papers,2015 +76,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,93420d9212dd15b3ef37f566e4d57e76bb2fab2f,citation,https://arxiv.org/pdf/1611.00851.pdf,An All-In-One Convolutional Neural Network for Face Analysis,2017 +77,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,def2983576001bac7d6461d78451159800938112,citation,https://arxiv.org/pdf/1705.07426.pdf,The Do’s and Don’ts for CNN-Based Face Verification,2017 +78,IJB-A,ijb_c,42.718568,-84.47791571,Michigan State University,edu,4b605e6a9362485bfe69950432fa1f896e7d19bf,citation,http://biometrics.cse.msu.edu/Publications/Face/BlantonAllenMillerKalkaJain_CVPRWB2016_HID.pdf,A Comparison of Human and Automated Face Verification Accuracy on Unconstrained Image Sets,2016 +79,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,8d3e95c31c93548b8c71dbeee2e9f7180067a888,citation,https://doi.org/10.1109/ICPR.2016.7899841,Template regularized sparse coding for face verification,2016 +80,IJB-A,ijb_c,42.8271556,-73.8780481,GE Global Research,company,8d3e95c31c93548b8c71dbeee2e9f7180067a888,citation,https://doi.org/10.1109/ICPR.2016.7899841,Template regularized sparse coding for face verification,2016 +81,IJB-A,ijb_c,25.0410728,121.6147562,Institute of Information Science,edu,337dd4aaca2c5f9b5d2de8e0e2401b5a8feb9958,citation,https://arxiv.org/pdf/1810.11160.pdf,Data-specific Adaptive Threshold for Face Recognition and Authentication,2018 +82,IJB-A,ijb_c,22.59805605,113.98533784,Shenzhen Institutes of Advanced Technology,edu,0aeb5020003e0c89219031b51bd30ff1bceea363,citation,http://doi.ieeecomputersociety.org/10.1109/CVPR.2016.525,Sparsifying Neural Network Connections for Face Recognition,2016 +83,IJB-A,ijb_c,22.42031295,114.20788644,Chinese University of Hong Kong,edu,0aeb5020003e0c89219031b51bd30ff1bceea363,citation,http://doi.ieeecomputersociety.org/10.1109/CVPR.2016.525,Sparsifying Neural Network Connections for Face Recognition,2016 +84,IJB-A,ijb_c,42.718568,-84.47791571,Michigan State University,edu,99daa2839213f904e279aec7cef26c1dfb768c43,citation,https://arxiv.org/pdf/1805.02283.pdf,DocFace: Matching ID Document Photos to Selfies,2018 +85,IJB-A,ijb_c,43.7776426,11.259765,University of Florence,edu,71ca8b6e84c17b3e68f980bfb8cddc837100f8bf,citation,http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7899774,Effective 3D based frontalization for unconstrained face recognition,2016 +86,IJB-A,ijb_c,51.49887085,-0.17560797,Imperial College London,edu,c43ed9b34cad1a3976bac7979808eb038d88af84,citation,https://arxiv.org/pdf/1804.03675.pdf,Semi-supervised Adversarial Learning to Generate Photorealistic Face Images of New Identities from 3D Morphable Model,2018 +87,IJB-A,ijb_c,51.24303255,-0.59001382,University of Surrey,edu,c43ed9b34cad1a3976bac7979808eb038d88af84,citation,https://arxiv.org/pdf/1804.03675.pdf,Semi-supervised Adversarial Learning to Generate Photorealistic Face Images of New Identities from 3D Morphable Model,2018 +88,IJB-A,ijb_c,37.3936717,-122.0807262,Facebook,company,628a3f027b7646f398c68a680add48c7969ab1d9,citation,https://pdfs.semanticscholar.org/628a/3f027b7646f398c68a680add48c7969ab1d9.pdf,Plan for Final Year Project : HKU-Face : A Large Scale Dataset for Deep Face Recognition,2017 +89,IJB-A,ijb_c,40.2773077,-7.5095801,University of Beira Interior,edu,61262450d4d814865a4f9a84299c24daa493f66e,citation,http://doi.org/10.1007/s10462-016-9474-x,Biometric recognition in surveillance scenarios: a survey,2016 +90,IJB-A,ijb_c,-31.95040445,115.79790037,University of Western Australia,edu,626913b8fcbbaee8932997d6c4a78fe1ce646127,citation,https://arxiv.org/pdf/1711.05942.pdf,Learning from Millions of 3D Scans for Large-scale 3D Face Recognition,2017 +91,IJB-A,ijb_c,35.9023226,14.4834189,University of Malta,edu,4efd58102ff46b7435c9ec6d4fc3dd21d93b15b4,citation,https://doi.org/10.1109/TIFS.2017.2788002,"Matching Software-Generated Sketches to Face Photographs With a Very Deep CNN, Morphed Faces, and Transfer Learning",2018 +92,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,b6f758be954d34817d4ebaa22b30c63a4b8ddb35,citation,http://arxiv.org/abs/1703.04835,A Proximity-Aware Hierarchical Clustering of Faces,2017 +93,IJB-A,ijb_c,32.77824165,34.99565673,Open University of Israel,edu,0a34fe39e9938ae8c813a81ae6d2d3a325600e5c,citation,https://arxiv.org/pdf/1708.07517.pdf,FacePoseNet: Making a Case for Landmark-Free Face Alignment,2017 +94,IJB-A,ijb_c,40.2773077,-7.5095801,University of Beira Interior,edu,84ae55603bffda40c225fe93029d39f04793e01f,citation,https://doi.org/10.1109/ICB.2016.7550066,ICB-RW 2016: International challenge on biometric recognition in the wild,2016 +95,IJB-A,ijb_c,41.70456775,-86.23822026,University of Notre Dame,edu,73ea06787925157df519a15ee01cc3dc1982a7e0,citation,https://arxiv.org/pdf/1811.01474.pdf,Fast Face Image Synthesis with Minimal Training,2018 +96,IJB-A,ijb_c,42.718568,-84.47791571,Michigan State University,edu,c6382de52636705be5898017f2f8ed7c70d7ae96,citation,http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7139089,Unconstrained face detection: State of the art baseline and challenges,2015 +97,IJB-A,ijb_c,38.95187,-77.363259,"Noblis, Falls Church, VA, U.S.A.",company,c6382de52636705be5898017f2f8ed7c70d7ae96,citation,http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7139089,Unconstrained face detection: State of the art baseline and challenges,2015 +98,IJB-A,ijb_c,40.47913175,-74.43168868,Rutgers University,edu,eee06d68497be8bf3a8aba4fde42a13aa090b301,citation,https://arxiv.org/pdf/1806.11191.pdf,CR-GAN: Learning Complete Representations for Multi-view Generation,2018 +99,IJB-A,ijb_c,35.3103441,-80.73261617,University of North Carolina at Charlotte,edu,eee06d68497be8bf3a8aba4fde42a13aa090b301,citation,https://arxiv.org/pdf/1806.11191.pdf,CR-GAN: Learning Complete Representations for Multi-view Generation,2018 +100,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,a3201e955d6607d383332f3a12a7befa08c5a18c,citation,http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7900276,VLAD encoded Deep Convolutional features for unconstrained face verification,2016 +101,IJB-A,ijb_c,40.47913175,-74.43168868,Rutgers University,edu,a3201e955d6607d383332f3a12a7befa08c5a18c,citation,http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7900276,VLAD encoded Deep Convolutional features for unconstrained face verification,2016 +102,IJB-A,ijb_c,22.42031295,114.20788644,Chinese University of Hong Kong,edu,52d7eb0fbc3522434c13cc247549f74bb9609c5d,citation,https://arxiv.org/pdf/1511.06523.pdf,WIDER FACE: A Face Detection Benchmark,2016 +103,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,19458454308a9f56b7de76bf7d8ff8eaa52b0173,citation,https://pdfs.semanticscholar.org/1945/8454308a9f56b7de76bf7d8ff8eaa52b0173.pdf,Deep Features for Recognizing Disguised Faces in the Wild,0 +104,IJB-A,ijb_c,43.7776426,11.259765,University of Florence,edu,746c0205fdf191a737df7af000eaec9409ede73f,citation,http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8423119,Investigating Nuisances in DCNN-Based Face Recognition,2018 +105,IJB-A,ijb_c,47.5612651,7.5752961,University of Basel,edu,0081e2188c8f34fcea3e23c49fb3e17883b33551,citation,http://pdfs.semanticscholar.org/0081/e2188c8f34fcea3e23c49fb3e17883b33551.pdf,Training Deep Face Recognition Systems with Synthetic Data,2018 +106,IJB-A,ijb_c,37.4102193,-122.05965487,Carnegie Mellon University,edu,2b869d5551b10f13bf6fcdb8d13f0aa4d1f59fc4,citation,https://arxiv.org/pdf/1803.00130.pdf,Ring loss: Convex Feature Normalization for Face Recognition,2018 +107,IJB-A,ijb_c,28.2290209,112.99483204,"National University of Defense Technology, China",edu,5f771fed91c8e4b666489ba2384d0705bcf75030,citation,https://arxiv.org/pdf/1804.03287.pdf,Understanding Humans in Crowded Scenes: Deep Nested Adversarial Learning and A New Benchmark for Multi-Human Parsing,2018 +108,IJB-A,ijb_c,1.2962018,103.77689944,National University of Singapore,edu,5f771fed91c8e4b666489ba2384d0705bcf75030,citation,https://arxiv.org/pdf/1804.03287.pdf,Understanding Humans in Crowded Scenes: Deep Nested Adversarial Learning and A New Benchmark for Multi-Human Parsing,2018 +109,IJB-A,ijb_c,42.3889785,-72.5286987,University of Massachusetts,edu,2241eda10b76efd84f3c05bdd836619b4a3df97e,citation,http://arxiv.org/pdf/1506.01342v5.pdf,One-to-many face recognition with bilinear CNNs,2016 +110,IJB-A,ijb_c,22.42031295,114.20788644,Chinese University of Hong Kong,edu,58d76380d194248b3bb291b8c7c5137a0a376897,citation,https://pdfs.semanticscholar.org/58d7/6380d194248b3bb291b8c7c5137a0a376897.pdf,FaceID-GAN : Learning a Symmetry Three-Player GAN for Identity-Preserving Face Synthesis,2018 +111,IJB-A,ijb_c,22.59805605,113.98533784,Shenzhen Institutes of Advanced Technology,edu,58d76380d194248b3bb291b8c7c5137a0a376897,citation,https://pdfs.semanticscholar.org/58d7/6380d194248b3bb291b8c7c5137a0a376897.pdf,FaceID-GAN : Learning a Symmetry Three-Player GAN for Identity-Preserving Face Synthesis,2018 +112,IJB-A,ijb_c,42.718568,-84.47791571,Michigan State University,edu,7fb5006b6522436ece5bedf509e79bdb7b79c9a7,citation,https://pdfs.semanticscholar.org/7fb5/006b6522436ece5bedf509e79bdb7b79c9a7.pdf,Multi-Task Convolutional Neural Network for Face Recognition,2017 +113,IJB-A,ijb_c,-27.49741805,153.01316956,University of Queensland,edu,28646c6220848db46c6944967298d89a6559c700,citation,https://pdfs.semanticscholar.org/2864/6c6220848db46c6944967298d89a6559c700.pdf,It takes two to tango : Cascading off-the-shelf face detectors,2018 +114,IJB-A,ijb_c,51.7534538,-1.25400997,University of Oxford,edu,5812d8239d691e99d4108396f8c26ec0619767a6,citation,https://arxiv.org/pdf/1810.09951.pdf,GhostVLAD for set-based face recognition,2018 +115,IJB-A,ijb_c,25.01353105,121.54173736,National Taiwan University of Science and Technology,edu,e4c3587392d477b7594086c6f28a00a826abf004,citation,https://doi.org/10.1109/ICIP.2017.8296998,Face recognition by facial attribute assisted network,2017 +116,IJB-A,ijb_c,1.3484104,103.68297965,Nanyang Technological University,edu,47190d213caef85e8b9dd0d271dbadc29ed0a953,citation,https://arxiv.org/pdf/1807.11649.pdf,The Devil of Face Recognition is in the Noise,2018 +117,IJB-A,ijb_c,32.87935255,-117.23110049,"University of California, San Diego",edu,47190d213caef85e8b9dd0d271dbadc29ed0a953,citation,https://arxiv.org/pdf/1807.11649.pdf,The Devil of Face Recognition is in the Noise,2018 +118,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,ce6d60b69eb95477596535227958109e07c61e1e,citation,http://www.rci.rutgers.edu/~vmp93/Conference_pub/BTAS_2015_FVFF_JunCheng_Chen.pdf,Unconstrained face verification using fisher vectors computed from frontalized faces,2015 +119,IJB-A,ijb_c,29.7207902,-95.34406271,University of Houston,edu,38d8ff137ff753f04689e6b76119a44588e143f3,citation,http://pdfs.semanticscholar.org/38d8/ff137ff753f04689e6b76119a44588e143f3.pdf,When 3D-Aided 2D Face Recognition Meets Deep Learning: An extended UR2D for Pose-Invariant Face Recognition,2017 +120,IJB-A,ijb_c,39.9082804,116.2458527,University of Chinese Academy of Sciences,edu,9627f28ea5f4c389350572b15968386d7ce3fe49,citation,https://arxiv.org/pdf/1802.07447.pdf,Load Balanced GANs for Multi-view Face Image Synthesis,2018 +121,IJB-A,ijb_c,34.0224149,-118.28634407,University of Southern California,edu,4e7ed13e541b8ed868480375785005d33530e06d,citation,http://doi.ieeecomputersociety.org/10.1109/WACV.2016.7477555,Face recognition using deep multi-pose representations,2016 +122,IJB-A,ijb_c,32.77824165,34.99565673,Open University of Israel,edu,582edc19f2b1ab2ac6883426f147196c8306685a,citation,http://pdfs.semanticscholar.org/be6c/db7b181e73f546d43cf2ab6bc7181d7d619b.pdf,Do We Really Need to Collect Millions of Faces for Effective Face Recognition?,2016 +123,IJB-A,ijb_c,37.4102193,-122.05965487,Carnegie Mellon University,edu,87e6cb090aecfc6f03a3b00650a5c5f475dfebe1,citation,https://pdfs.semanticscholar.org/87e6/cb090aecfc6f03a3b00650a5c5f475dfebe1.pdf,Holistically Constrained Local Model: Going Beyond Frontal Poses for Facial Landmark Detection,2016 +124,IJB-A,ijb_c,34.0224149,-118.28634407,University of Southern California,edu,87e6cb090aecfc6f03a3b00650a5c5f475dfebe1,citation,https://pdfs.semanticscholar.org/87e6/cb090aecfc6f03a3b00650a5c5f475dfebe1.pdf,Holistically Constrained Local Model: Going Beyond Frontal Poses for Facial Landmark Detection,2016 +125,IJB-A,ijb_c,39.65404635,-79.96475355,West Virginia University,edu,3b9b200e76a35178da940279d566bbb7dfebb787,citation,http://pdfs.semanticscholar.org/3b9b/200e76a35178da940279d566bbb7dfebb787.pdf,Learning Channel Inter-dependencies at Multiple Scales on Dense Networks for Face Recognition,2017 +126,IJB-A,ijb_c,-27.49741805,153.01316956,University of Queensland,edu,de79437f74e8e3b266afc664decf4e6e4bdf34d7,citation,https://doi.org/10.1109/IVCNZ.2016.7804415,To face or not to face: Towards reducing false positive of face detection,2016 +127,IJB-A,ijb_c,46.0501558,14.46907327,University of Ljubljana,edu,368d59cf1733af511ed8abbcbeb4fb47afd4da1c,citation,http://pdfs.semanticscholar.org/368d/59cf1733af511ed8abbcbeb4fb47afd4da1c.pdf,To Frontalize or Not To Frontalize: A Study of Face Pre-Processing Techniques and Their Impact on Recognition,2016 +128,IJB-A,ijb_c,41.70456775,-86.23822026,University of Notre Dame,edu,368d59cf1733af511ed8abbcbeb4fb47afd4da1c,citation,http://pdfs.semanticscholar.org/368d/59cf1733af511ed8abbcbeb4fb47afd4da1c.pdf,To Frontalize or Not To Frontalize: A Study of Face Pre-Processing Techniques and Their Impact on Recognition,2016 +129,IJB-A,ijb_c,32.77824165,34.99565673,Open University of Israel,edu,62e913431bcef5983955e9ca160b91bb19d9de42,citation,http://pdfs.semanticscholar.org/62e9/13431bcef5983955e9ca160b91bb19d9de42.pdf,Facial Landmark Detection with Tweaked Convolutional Neural Networks,2015 +130,IJB-A,ijb_c,29.5084174,106.57858552,Chongqing University,edu,acd4280453b995cb071c33f7c9db5760432f4279,citation,https://doi.org/10.1007/s00138-018-0907-1,Deep transformation learning for face recognition in the unconstrained scene,2018 +131,IJB-A,ijb_c,38.99203005,-76.9461029,University of Maryland College Park,edu,ceeb67bf53ffab1395c36f1141b516f893bada27,citation,http://pdfs.semanticscholar.org/ceeb/67bf53ffab1395c36f1141b516f893bada27.pdf,Face Alignment by Local Deep Descriptor Regression,2016 +132,IJB-A,ijb_c,40.47913175,-74.43168868,Rutgers University,edu,ceeb67bf53ffab1395c36f1141b516f893bada27,citation,http://pdfs.semanticscholar.org/ceeb/67bf53ffab1395c36f1141b516f893bada27.pdf,Face Alignment by Local Deep Descriptor Regression,2016 +133,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,ceeb67bf53ffab1395c36f1141b516f893bada27,citation,http://pdfs.semanticscholar.org/ceeb/67bf53ffab1395c36f1141b516f893bada27.pdf,Face Alignment by Local Deep Descriptor Regression,2016 +134,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,37619564574856c6184005830deda4310d3ca580,citation,https://doi.org/10.1109/BTAS.2015.7358755,A deep pyramid Deformable Part Model for face detection,2015 +135,IJB-A,ijb_c,51.7534538,-1.25400997,University of Oxford,edu,eb027969f9310e0ae941e2adee2d42cdf07d938c,citation,https://arxiv.org/pdf/1710.08092.pdf,VGGFace2: A Dataset for Recognising Faces across Pose and Age,2018 +136,IJB-A,ijb_c,42.3889785,-72.5286987,University of Massachusetts,edu,3c97c32ff575989ef2869f86d89c63005fc11ba9,citation,http://people.cs.umass.edu/~hzjiang/pubs/face_det_fg_2017.pdf,Face Detection with the Faster R-CNN,2017 +137,IJB-A,ijb_c,39.2899685,-76.62196103,University of Maryland,edu,4f7b92bd678772552b3c3edfc9a7c5c4a8c60a8e,citation,https://pdfs.semanticscholar.org/4f7b/92bd678772552b3c3edfc9a7c5c4a8c60a8e.pdf,Deep Density Clustering of Unconstrained Faces,0 +138,IJB-A,ijb_c,1.2962018,103.77689944,National University of Singapore,edu,fca9ebaa30d69ccec8bb577c31693c936c869e72,citation,https://arxiv.org/pdf/1809.00338.pdf,Look Across Elapse: Disentangled Representation Learning and Photorealistic Cross-Age Face Synthesis for Age-Invariant Face Recognition,2018 +139,IJB-A,ijb_c,40.0044795,116.370238,Chinese Academy of Sciences,edu,fca9ebaa30d69ccec8bb577c31693c936c869e72,citation,https://arxiv.org/pdf/1809.00338.pdf,Look Across Elapse: Disentangled Representation Learning and Photorealistic Cross-Age Face Synthesis for Age-Invariant Face Recognition,2018 -- cgit v1.2.3-70-g09d2 From dca2b5054b01fd2d1ffe09dde161b200c95b2a59 Mon Sep 17 00:00:00 2001 From: Adam Harvey Date: Thu, 28 Feb 2019 15:52:52 +0100 Subject: add images --- .../50_people_one_question/assets/background.jpg | Bin 0 -> 41129 bytes .../assets/background_02.jpg | Bin 0 -> 51059 bytes .../50_people_one_question/assets/index.jpg | Bin 2981 -> 16384 bytes .../50_people_one_question/assets/index_02.jpg | Bin 3064 -> 10995 bytes .../pages/datasets/brainwash/assets/background.jpg | Bin 0 -> 35910 bytes .../pages/datasets/brainwash/assets/index.jpg | Bin 3056 -> 9921 bytes .../pages/datasets/duke_mtmc/assets/background.jpg | Bin 0 -> 28688 bytes .../pages/datasets/duke_mtmc/assets/index.jpg | Bin 2436 -> 8519 bytes .../pages/datasets/helen/assets/background.jpg | Bin 0 -> 11553 bytes site/content/pages/datasets/helen/assets/index.jpg | Bin 3243 -> 7556 bytes .../datasets/hrt_transgender/assets/background.jpg | Bin 0 -> 11727 bytes .../datasets/hrt_transgender/assets/index.jpg | Bin 3085 -> 7439 bytes .../pages/datasets/ilids_vid/assets/background.jpg | Bin 0 -> 245031 bytes .../pages/datasets/ilids_vid/assets/index.jpg | Bin 0 -> 22313 bytes .../pages/datasets/lfw/assets/background.jpg | Bin 239321 -> 316873 bytes site/content/pages/datasets/lfw/assets/index.jpg | Bin 25306 -> 25757 bytes .../pages/datasets/mars/assets/background.jpg | Bin 216396 -> 308757 bytes site/content/pages/datasets/mars/assets/index.jpg | Bin 23722 -> 24177 bytes .../pages/datasets/pubfig/assets/background.jpg | Bin 0 -> 159672 bytes .../content/pages/datasets/pubfig/assets/index.jpg | Bin 20533 -> 20802 bytes .../pages/datasets/uccs/assets/background.jpg | Bin 0 -> 37310 bytes site/content/pages/datasets/uccs/assets/index.jpg | Bin 0 -> 9609 bytes .../datasets/ucf_selfie/assets/background.jpg | Bin 0 -> 9241 bytes .../datasets/ucf_selfie/assets/background_02.jpg | Bin 0 -> 8439 bytes .../pages/datasets/ucf_selfie/assets/index.jpg | Bin 0 -> 6241 bytes .../pages/datasets/ucf_selfie/assets/index_02.jpg | Bin 0 -> 4937 bytes site/content/pages/datasets/ytmu/assets/index.jpg | Bin 6489 -> 5936 bytes .../pages/datasets/ytmu/assets/index_04.jpg | Bin 0 -> 9789 bytes 28 files changed, 0 insertions(+), 0 deletions(-) create mode 100644 site/content/pages/datasets/50_people_one_question/assets/background.jpg create mode 100644 site/content/pages/datasets/50_people_one_question/assets/background_02.jpg create mode 100644 site/content/pages/datasets/brainwash/assets/background.jpg create mode 100644 site/content/pages/datasets/duke_mtmc/assets/background.jpg create mode 100644 site/content/pages/datasets/helen/assets/background.jpg create mode 100644 site/content/pages/datasets/hrt_transgender/assets/background.jpg create mode 100644 site/content/pages/datasets/ilids_vid/assets/background.jpg create mode 100644 site/content/pages/datasets/ilids_vid/assets/index.jpg create mode 100644 site/content/pages/datasets/pubfig/assets/background.jpg create mode 100644 site/content/pages/datasets/uccs/assets/background.jpg create mode 100644 site/content/pages/datasets/uccs/assets/index.jpg create mode 100644 site/content/pages/datasets/ucf_selfie/assets/background.jpg create mode 100644 site/content/pages/datasets/ucf_selfie/assets/background_02.jpg create mode 100644 site/content/pages/datasets/ucf_selfie/assets/index.jpg create mode 100644 site/content/pages/datasets/ucf_selfie/assets/index_02.jpg create mode 100644 site/content/pages/datasets/ytmu/assets/index_04.jpg (limited to 'site') diff --git a/site/content/pages/datasets/50_people_one_question/assets/background.jpg b/site/content/pages/datasets/50_people_one_question/assets/background.jpg new file mode 100644 index 00000000..f7567314 Binary files /dev/null and b/site/content/pages/datasets/50_people_one_question/assets/background.jpg differ diff --git a/site/content/pages/datasets/50_people_one_question/assets/background_02.jpg b/site/content/pages/datasets/50_people_one_question/assets/background_02.jpg new file mode 100644 index 00000000..f20ee861 Binary files /dev/null and b/site/content/pages/datasets/50_people_one_question/assets/background_02.jpg differ diff --git a/site/content/pages/datasets/50_people_one_question/assets/index.jpg b/site/content/pages/datasets/50_people_one_question/assets/index.jpg index a79c7739..7bf4ed1d 100644 Binary files a/site/content/pages/datasets/50_people_one_question/assets/index.jpg and b/site/content/pages/datasets/50_people_one_question/assets/index.jpg differ diff --git a/site/content/pages/datasets/50_people_one_question/assets/index_02.jpg b/site/content/pages/datasets/50_people_one_question/assets/index_02.jpg index c331ea5c..347bd08d 100644 Binary files a/site/content/pages/datasets/50_people_one_question/assets/index_02.jpg and b/site/content/pages/datasets/50_people_one_question/assets/index_02.jpg differ diff --git a/site/content/pages/datasets/brainwash/assets/background.jpg b/site/content/pages/datasets/brainwash/assets/background.jpg new file mode 100644 index 00000000..eada1779 Binary files /dev/null and b/site/content/pages/datasets/brainwash/assets/background.jpg differ diff --git a/site/content/pages/datasets/brainwash/assets/index.jpg b/site/content/pages/datasets/brainwash/assets/index.jpg index 7d6230e1..c903baea 100644 Binary files a/site/content/pages/datasets/brainwash/assets/index.jpg and b/site/content/pages/datasets/brainwash/assets/index.jpg differ diff --git a/site/content/pages/datasets/duke_mtmc/assets/background.jpg b/site/content/pages/datasets/duke_mtmc/assets/background.jpg new file mode 100644 index 00000000..fb76d97f Binary files /dev/null and b/site/content/pages/datasets/duke_mtmc/assets/background.jpg differ diff --git a/site/content/pages/datasets/duke_mtmc/assets/index.jpg b/site/content/pages/datasets/duke_mtmc/assets/index.jpg index 6651c15c..89b7b08d 100644 Binary files a/site/content/pages/datasets/duke_mtmc/assets/index.jpg and b/site/content/pages/datasets/duke_mtmc/assets/index.jpg differ diff --git a/site/content/pages/datasets/helen/assets/background.jpg b/site/content/pages/datasets/helen/assets/background.jpg new file mode 100644 index 00000000..63ebb8c9 Binary files /dev/null and b/site/content/pages/datasets/helen/assets/background.jpg differ diff --git a/site/content/pages/datasets/helen/assets/index.jpg b/site/content/pages/datasets/helen/assets/index.jpg index 37cb5882..9e62679b 100644 Binary files a/site/content/pages/datasets/helen/assets/index.jpg and b/site/content/pages/datasets/helen/assets/index.jpg differ diff --git a/site/content/pages/datasets/hrt_transgender/assets/background.jpg b/site/content/pages/datasets/hrt_transgender/assets/background.jpg new file mode 100644 index 00000000..9437c4ff Binary files /dev/null and b/site/content/pages/datasets/hrt_transgender/assets/background.jpg differ diff --git a/site/content/pages/datasets/hrt_transgender/assets/index.jpg b/site/content/pages/datasets/hrt_transgender/assets/index.jpg index 4f3307c4..428ac42a 100644 Binary files a/site/content/pages/datasets/hrt_transgender/assets/index.jpg and b/site/content/pages/datasets/hrt_transgender/assets/index.jpg differ diff --git a/site/content/pages/datasets/ilids_vid/assets/background.jpg b/site/content/pages/datasets/ilids_vid/assets/background.jpg new file mode 100644 index 00000000..6acd30e8 Binary files /dev/null and b/site/content/pages/datasets/ilids_vid/assets/background.jpg differ diff --git a/site/content/pages/datasets/ilids_vid/assets/index.jpg b/site/content/pages/datasets/ilids_vid/assets/index.jpg new file mode 100644 index 00000000..66c80d7d Binary files /dev/null and b/site/content/pages/datasets/ilids_vid/assets/index.jpg differ diff --git a/site/content/pages/datasets/lfw/assets/background.jpg b/site/content/pages/datasets/lfw/assets/background.jpg index 64d61c35..3ab1607d 100644 Binary files a/site/content/pages/datasets/lfw/assets/background.jpg and b/site/content/pages/datasets/lfw/assets/background.jpg differ diff --git a/site/content/pages/datasets/lfw/assets/index.jpg b/site/content/pages/datasets/lfw/assets/index.jpg index 1df0c4f6..bc36c106 100644 Binary files a/site/content/pages/datasets/lfw/assets/index.jpg and b/site/content/pages/datasets/lfw/assets/index.jpg differ diff --git a/site/content/pages/datasets/mars/assets/background.jpg b/site/content/pages/datasets/mars/assets/background.jpg index 9c16c26d..f3440590 100644 Binary files a/site/content/pages/datasets/mars/assets/background.jpg and b/site/content/pages/datasets/mars/assets/background.jpg differ diff --git a/site/content/pages/datasets/mars/assets/index.jpg b/site/content/pages/datasets/mars/assets/index.jpg index 9e527fbd..e866defd 100644 Binary files a/site/content/pages/datasets/mars/assets/index.jpg and b/site/content/pages/datasets/mars/assets/index.jpg differ diff --git a/site/content/pages/datasets/pubfig/assets/background.jpg b/site/content/pages/datasets/pubfig/assets/background.jpg new file mode 100644 index 00000000..db748a8f Binary files /dev/null and b/site/content/pages/datasets/pubfig/assets/background.jpg differ diff --git a/site/content/pages/datasets/pubfig/assets/index.jpg b/site/content/pages/datasets/pubfig/assets/index.jpg index be14f27c..2470b35c 100644 Binary files a/site/content/pages/datasets/pubfig/assets/index.jpg and b/site/content/pages/datasets/pubfig/assets/index.jpg differ diff --git a/site/content/pages/datasets/uccs/assets/background.jpg b/site/content/pages/datasets/uccs/assets/background.jpg new file mode 100644 index 00000000..db6b0180 Binary files /dev/null and b/site/content/pages/datasets/uccs/assets/background.jpg differ diff --git a/site/content/pages/datasets/uccs/assets/index.jpg b/site/content/pages/datasets/uccs/assets/index.jpg new file mode 100644 index 00000000..9809e564 Binary files /dev/null and b/site/content/pages/datasets/uccs/assets/index.jpg differ diff --git a/site/content/pages/datasets/ucf_selfie/assets/background.jpg b/site/content/pages/datasets/ucf_selfie/assets/background.jpg new file mode 100644 index 00000000..877133e5 Binary files /dev/null and b/site/content/pages/datasets/ucf_selfie/assets/background.jpg differ diff --git a/site/content/pages/datasets/ucf_selfie/assets/background_02.jpg b/site/content/pages/datasets/ucf_selfie/assets/background_02.jpg new file mode 100644 index 00000000..1e092f61 Binary files /dev/null and b/site/content/pages/datasets/ucf_selfie/assets/background_02.jpg differ diff --git a/site/content/pages/datasets/ucf_selfie/assets/index.jpg b/site/content/pages/datasets/ucf_selfie/assets/index.jpg new file mode 100644 index 00000000..4837f6e3 Binary files /dev/null and b/site/content/pages/datasets/ucf_selfie/assets/index.jpg differ diff --git a/site/content/pages/datasets/ucf_selfie/assets/index_02.jpg b/site/content/pages/datasets/ucf_selfie/assets/index_02.jpg new file mode 100644 index 00000000..524340d7 Binary files /dev/null and b/site/content/pages/datasets/ucf_selfie/assets/index_02.jpg differ diff --git a/site/content/pages/datasets/ytmu/assets/index.jpg b/site/content/pages/datasets/ytmu/assets/index.jpg index 6df15db5..76716847 100644 Binary files a/site/content/pages/datasets/ytmu/assets/index.jpg and b/site/content/pages/datasets/ytmu/assets/index.jpg differ diff --git a/site/content/pages/datasets/ytmu/assets/index_04.jpg b/site/content/pages/datasets/ytmu/assets/index_04.jpg new file mode 100644 index 00000000..3d67baac Binary files /dev/null and b/site/content/pages/datasets/ytmu/assets/index_04.jpg differ -- cgit v1.2.3-70-g09d2 From 6711fb0c58e969284e3fcf94bb163c77445e2e13 Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Thu, 28 Feb 2019 15:56:04 +0100 Subject: footnote back and forth navigation --- client/util/index.js | 4 ++ megapixels/app/site/parser.py | 17 +++++++- site/assets/css/css.css | 66 +++++++++++++++++++++++--------- site/content/pages/datasets/lfw/index.md | 2 +- site/public/datasets/lfw/index.html | 18 ++++----- 5 files changed, 77 insertions(+), 30 deletions(-) (limited to 'site') diff --git a/client/util/index.js b/client/util/index.js index d0db0d98..0792e24e 100644 --- a/client/util/index.js +++ b/client/util/index.js @@ -5,12 +5,16 @@ export const isiPad = !!(navigator.userAgent.match(/iPad/i)) export const isAndroid = !!(navigator.userAgent.match(/Android/i)) export const isMobile = isiPhone || isiPad || isAndroid export const isDesktop = !isMobile +export const isFirefox = typeof InstallTrigger !== 'undefined' export const toArray = a => Array.prototype.slice.apply(a) export const choice = a => a[Math.floor(Math.random() * a.length)] const htmlClassList = document.body.parentNode.classList htmlClassList.add(isDesktop ? 'desktop' : 'mobile') +if (isFirefox) { + htmlClassList.add('firefox') +} /* Default image dimensions */ diff --git a/megapixels/app/site/parser.py b/megapixels/app/site/parser.py index ef83b655..9e904e00 100644 --- a/megapixels/app/site/parser.py +++ b/megapixels/app/site/parser.py @@ -10,6 +10,8 @@ import app.site.s3 as s3 renderer = mistune.Renderer(escape=False) markdown = mistune.Markdown(renderer=renderer) +footnote_count = 0 + def parse_markdown(metadata, sections, s3_path, skip_h1=False): """ parse page into sections, preprocess the markdown to handle our modifications @@ -94,7 +96,18 @@ def parse_markdown(metadata, sections, s3_path, skip_h1=False): if footnote_lookup: for key, index in footnote_lookup.items(): - content = content.replace(key, '{}'.format(key, index, index)) + global footnote_count + footnote_count = 0 + letters = "abcdefghijklmnopqrstuvwxyz" + footnote_backlinks = [] + def footnote_tag(match): + global footnote_count + footnote_count += 1 + footnote_backlinks.append('{}'.format(key, footnote_count, letters[footnote_count-1])) + return ' {}'.format(key, footnote_count, key, index, index) + key_regex = re.compile(key.replace('[', '\\[').replace('^', '\\^').replace(']', '\\]')) + content = key_regex.sub(footnote_tag, content) + footnote_txt = footnote_txt.replace("{}_BACKLINKS".format(index), "".join(footnote_backlinks)) content += footnote_txt return content @@ -197,7 +210,7 @@ def format_footnotes(footnotes, s3_path): continue key, note = footnote.split(': ', 1) footnote_index_lookup[key] = index - footnote_list.append('^'.format(key) + markdown(note)) + footnote_list.append('{}_BACKLINKS'.format(key, index) + markdown(note)) index += 1 footnote_txt = '
          • ' + '
          • '.join(footnote_list) + '
          ' diff --git a/site/assets/css/css.css b/site/assets/css/css.css index 0afa3725..4b42657b 100644 --- a/site/assets/css/css.css +++ b/site/assets/css/css.css @@ -16,7 +16,8 @@ html { opacity: 0; transition: opacity 0.2s cubic-bezier(0,1,1,1); } -html.desktop .content, html.mobile .content { +html.desktop .content, +html.mobile .content { opacity: 1; } @@ -28,7 +29,7 @@ header { left: 0; width: 100%; height: 70px; - z-index: 2; + z-index: 9999; background: #1e1e1e; display: flex; flex-direction: row; @@ -53,8 +54,10 @@ header .logo { height: 30px; } header .site_name { + font-family: 'Roboto', sans-serif; font-weight: bold; color: #fff; + font-size: 14px; } header .sub { margin-left: 4px; @@ -148,7 +151,7 @@ h3 { margin: 0 0 20px 0; padding: 0; font-size: 14pt; - font-weight: 600; + font-weight: 500; transition: color 0.2s cubic-bezier(0,0,1,1); } h4 { @@ -170,6 +173,8 @@ h4 { margin: 0; padding: 0 0 10px 0; font-family: 'Roboto Mono'; + font-weight: 400; + font-size: 11px; text-transform: uppercase; letter-spacing: 2px; } @@ -210,13 +215,17 @@ section { p { margin: 0 0 20px 0; line-height: 2; + font-size: 15px; + font-weight: 400; } .content a { - color: #ff0; + color: #fff; + text-decoration: none; + border-bottom: 1px dashed; transition: color 0.2s cubic-bezier(0,0,1,1); } -.content a:hover { - color: #fff; +.desktop .content a:hover { + color: #ff8; } /* top of post metadata */ @@ -368,7 +377,7 @@ section.fullwidth .image { .caption { text-align: left; font-size: 9pt; - color: #bbb; + color: #999; max-width: 960px; margin: 10px auto 0 auto; font-family: 'Roboto'; @@ -538,17 +547,22 @@ section.intro_section { font-size: 38px; line-height: 60px; margin-bottom: 30px; - color: #fff; + color: #ddd; + font-weight: 300; } .intro_section .hero_subdesc { font-size: 18px; line-height: 36px; max-width: 640px; + font-weight: 300; color: #ddd; } -.intro_section span { - box-shadow: -10px -10px #000, 10px -10px #000, 10px 10px #000, -10px 10px #000; - background: #000; +.intro_section div > span { + box-shadow: -10px -10px #1e1e1e, 10px -10px #1e1e1e, 10px 10px #1e1e1e, -10px 10px #1e1e1e; + background: #1e1e1e; +} +.firefox .intro_section div > span { + box-decoration-break: clone; } /* footnotes */ @@ -559,22 +573,38 @@ a.footnote { display: inline-block; bottom: 10px; text-decoration: none; - color: #ff0; + color: #ff8; + border: 0; left: 2px; + transition-duration: 0s; +} +a.footnote_shim { + display: inline-block; + width: 1px; height: 1px; + overflow: hidden; + position: relative; + top: -90px; + visibility: hidden; } .right-sidebar a.footnote { bottom: 8px; } .desktop a.footnote:hover { - background-color: #ff0; + background-color: #ff8; color: #000; } -a.footnote_anchor { - font-weight: bold; - color: #ff0; +.backlinks { margin-right: 10px; - text-decoration: underline; - cursor: pointer; +} +.content .backlinks a { + color: #ff8; + font-size: 10px; + text-decoration: none; + border: 0; + font-weight: bold; + position: relative; + bottom: 5px; + margin-right: 2px; } ul.footnotes { list-style-type: decimal; diff --git a/site/content/pages/datasets/lfw/index.md b/site/content/pages/datasets/lfw/index.md index 1995e1f9..972fafe2 100644 --- a/site/content/pages/datasets/lfw/index.md +++ b/site/content/pages/datasets/lfw/index.md @@ -5,7 +5,7 @@ title: Labeled Faces in The Wild desc: Labeled Faces in The Wild (LFW) is a database of face photographs designed for studying the problem of unconstrained face recognition. subdesc: It includes 13,456 images of 4,432 people’s images copied from the Internet during 2002-2004. image: assets/lfw_feature.jpg -caption: Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms. +caption: A few of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms. slug: lfw published: 2019-2-23 updated: 2019-2-23 diff --git a/site/public/datasets/lfw/index.html b/site/public/datasets/lfw/index.html index 54b6aa22..08ec8ee3 100644 --- a/site/public/datasets/lfw/index.html +++ b/site/public/datasets/lfw/index.html @@ -28,10 +28,10 @@
          Labeled Faces in The Wild (LFW) is a database of face photographs designed for studying the problem of unconstrained face recognition.
          It includes 13,456 images of 4,432 people’s images copied from the Internet during 2002-2004. -
          Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.

          Statistics

          +
          A few of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.

          Statistics

          Years
          2002-2004
          Images
          13,233
          Identities
          5,749
          Origin
          Yahoo News Images
          Funding
          (Possibly, partially CIA)

          INSIGHTS

            -
          • There are about 3 men for every 1 woman (4,277 men and 1,472 women) in the LFW dataset1
            • +
          • There are about 3 men for every 1 woman (4,277 men and 1,472 women) in the LFW dataset 1
          • The person with the most images is George W. Bush with 530
          • There are about 3 George W. Bush's for every 1 Tony Blair
          • The LFW dataset includes over 500 actors, 30 models, 10 presidents, 124 basketball players, 24 football players, 11 kings, 7 queens, and 1 Moby
            • @@ -39,7 +39,7 @@
          • The word "future" appears 71 times

          Labeled Faces in the Wild

          -

          Labeled Faces in The Wild (LFW) is "a database of face photographs designed for studying the problem of unconstrained face recognition1. It is used to evaluate and improve the performance of facial recognition algorithms in academic, commercial, and government research. According to BiometricUpdate.com3, LFW is "the most widely used evaluation set in the field of facial recognition, LFW attracts a few dozen teams from around the globe including Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong."

          +

          Labeled Faces in The Wild (LFW) is "a database of face photographs designed for studying the problem of unconstrained face recognition 1. It is used to evaluate and improve the performance of facial recognition algorithms in academic, commercial, and government research. According to BiometricUpdate.com 3, LFW is "the most widely used evaluation set in the field of facial recognition, LFW attracts a few dozen teams from around the globe including Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong."

          The LFW dataset includes 13,233 images of 5,749 people that were collected between 2002-2004. LFW is a subset of Names of Faces and is part of the first facial recognition training dataset created entirely from images appearing on the Internet. The people appearing in LFW are...

          The Names and Faces dataset was the first face recognition dataset created entire from online photos. However, Names and Faces and LFW are not the first face recognition dataset created entirely "in the wild". That title belongs to the UCD dataset. Images obtained "in the wild" means using an image without explicit consent or awareness from the subject or photographer.

          Biometric Trade Routes

          @@ -51,11 +51,11 @@

          Additional Information

          (tweet-sized snippets go here)

            -
          • The LFW dataset is considered the "most popular benchmark for face recognition" 2
            • -
          • The LFW dataset is "the most widely used evaluation set in the field of facial recognition" 3
            • +
          • The LFW dataset is considered the "most popular benchmark for face recognition" 2
            • +
          • The LFW dataset is "the most widely used evaluation set in the field of facial recognition" 3
          • All images in LFW dataset were obtained "in the wild" meaning without any consent from the subject or from the photographer
          • The faces in the LFW dataset were detected using the Viola-Jones haarcascade face detector [^lfw_website] [^lfw-survey]
            • -
          • The LFW dataset is used by several of the largest tech companies in the world including "Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong." 3
            • +
          • The LFW dataset is used by several of the largest tech companies in the world including "Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++ and Chinese University of Hong Kong." 3
          • All images in the LFW dataset were copied from Yahoo News between 2002 - 2004
          • In 2014, two of the four original authors of the LFW dataset received funding from IARPA and ODNI for their followup paper Labeled Faces in the Wild: Updates and New Reporting Procedures via IARPA contract number 2014-14071600010
          • The dataset includes 2 images of George Tenet, the former Director of Central Intelligence (DCI) for the Central Intelligence Agency whose facial biometrics were eventually used to help train facial recognition software in China and Russia
            • @@ -94,9 +94,9 @@ imageio.imwrite('lfw_montage_960.jpg', montage)

          Supplementary Material

          Text and graphics ©Adam Harvey / megapixels.cc

          -
          -- cgit v1.2.3-70-g09d2 From e9c404c06f47a0ad6c2a2775795a58b3f3d3a160 Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Thu, 28 Feb 2019 16:03:24 +0100 Subject: css goodies --- site/assets/css/applets.css | 3 +++ site/assets/css/css.css | 11 ++++++----- 2 files changed, 9 insertions(+), 5 deletions(-) (limited to 'site') diff --git a/site/assets/css/applets.css b/site/assets/css/applets.css index aa9ce47f..729737fe 100644 --- a/site/assets/css/applets.css +++ b/site/assets/css/applets.css @@ -143,6 +143,9 @@ /* tabulator */ +.tabulator { + font-family: 'Roboto', sans-serif; +} .tabulator-row { transition: background-color 100ms cubic-bezier(0,0,1,1); background-color: rgba(255,255,255,0.0); diff --git a/site/assets/css/css.css b/site/assets/css/css.css index 4b42657b..d710b3a8 100644 --- a/site/assets/css/css.css +++ b/site/assets/css/css.css @@ -9,7 +9,7 @@ html, body { overflow-x: hidden; } html { - background: #111111; + background: #181818; } .content { @@ -149,7 +149,7 @@ h2 { } h3 { margin: 0 0 20px 0; - padding: 0; + padding: 20px 0 0 0; font-size: 14pt; font-weight: 500; transition: color 0.2s cubic-bezier(0,0,1,1); @@ -174,7 +174,7 @@ h4 { padding: 0 0 10px 0; font-family: 'Roboto Mono'; font-weight: 400; - font-size: 11px; + font-size: 14px; text-transform: uppercase; letter-spacing: 2px; } @@ -263,6 +263,7 @@ p { } .right-sidebar ul { margin-bottom: 10px; + color: #aaa; } /* lists */ @@ -558,8 +559,8 @@ section.intro_section { color: #ddd; } .intro_section div > span { - box-shadow: -10px -10px #1e1e1e, 10px -10px #1e1e1e, 10px 10px #1e1e1e, -10px 10px #1e1e1e; - background: #1e1e1e; + box-shadow: -10px -10px #181818, 10px -10px #181818, 10px 10px #181818, -10px 10px #181818; + background: #181818; } .firefox .intro_section div > span { box-decoration-break: clone; -- cgit v1.2.3-70-g09d2 From d5cc74fd0805f67237a1065cd667e05f6b3616d9 Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Thu, 28 Feb 2019 16:07:05 +0100 Subject: comments --- README.md | 36 ++++++++++++++-------- site/public/research/00_introduction/index.html | 7 +++-- .../research/01_from_1_to_100_pixels/index.html | 3 ++ 3 files changed, 31 insertions(+), 15 deletions(-) (limited to 'site') diff --git a/README.md b/README.md index e46a6289..27dd1b38 100644 --- a/README.md +++ b/README.md @@ -38,28 +38,38 @@ You may need to set the database charset to `utf8mb4` in order to import the CSV ALTER DATABASE megapixels CHARACTER SET = utf8mb4 COLLATE = utf8mb4_unicode_ci; ``` -## Building the site +## Development: automatic rebuilds -The most recently built copy of the site is kept in the repo. This is generated directly from NextCloud. Be mindful that NextCloud will create extra copies of things if there are merge conflicts. +In development, we can watch a bunch of things and rebuild stuff automatically. These rebuild the HTML and the Javascript: ``` -npm install -npm run build -cd megapixels -python cli_faiss.py sync_metadata -python cli_faiss.py build_faiss -python cli_faiss.py build_db -python cli_site.py build +python cli_site.py watch +npm run watch ``` -## Running the site - -On OSX, you must run the server with `pythonw` because of matplotlib. +In addition, run the server, which will serve some HTML (you may need to add index.html to URLs... alas): ``` python cli_flask.py run +``` + +These servers must be running to use all features of the site (face search, etc.) + +``` python `which celery` worker -A app.server.tasks --loglevel=info -E redis-server /usr/local/etc/redis.conf -npm run watch ``` +Note: On OSX, you must run the server with `pythonw` because of matplotlib. + +## Building the site for production + +``` +npm install +npm run build +cd megapixels +python cli_faiss.py sync_metadata +python cli_faiss.py build_faiss +python cli_faiss.py build_db +python cli_site.py build +``` diff --git a/site/public/research/00_introduction/index.html b/site/public/research/00_introduction/index.html index 64047134..395bd268 100644 --- a/site/public/research/00_introduction/index.html +++ b/site/public/research/00_introduction/index.html @@ -42,8 +42,11 @@ -
          Posted
          Dec. 15
          Author
          Adam Harvey

          Ignore content below these lines

          -

          Ever since the first computational facial recognition research project by the CIA in the early 1960s, data has always played a vital role in the development of our biometric future. Without facial recognition datasets there would be no facial recognition. Datasets are an indispensable part of any artificial intelligence system because, as Geoffrey Hinton points out, "we no longer program computers with code, we program them with data".

          +
          Posted
          Dec. 15
          Author
          Adam Harvey

          Ever since the first computational facial recognition research project by the CIA in the early 1960s, data has always played a vital role in the development of our biometric future. Without facial recognition datasets there would be no facial recognition. Datasets are an indispensable part of any artificial intelligence system because, as Geoffrey Hinton points out:

          +

          Our relationship to computers has changed. Instead of programming them, we now show them and they figure it out. - Geoffrey Hinton

          +
          +

          Algorithms learn from datasets. And we program algorithms by building datasets. But datasets aren't like code. There's no programming language made of data except for the data itself.

          +

          Ignore content below these lines

          It was the early 2000s. Face recognition was new and no one seemed sure exactly how well it was going to perform in practice. In theory, face recognition was poised to be a game changer, a force multiplier, a strategic military advantage, a way to make cities safer and to secure borders. This was the future John Ashcroft demanded with the Total Information Awareness act of the 2003 and that spooks had dreamed of for decades. It was a future that academics at Carnegie Mellon Universtiy and Colorado State University would help build. It was also a future that celebrities would play a significant role in building. And to the surprise of ordinary Internet users like myself and perhaps you, it was a future that millions of Internet users would unwittingly play role in creating.

          Now the future has arrived and it doesn't make sense. Facial recognition works yet it doesn't actually work. Facial recognition is cheap and accessible but also expensive and out of control. Facial recognition research has achieved headline grabbing superhuman accuracies over 99.9% yet facial recognition is also dangerously inaccurate. During a trial installation at Sudkreuz station in Berlin in 2018, 20% of the matches were wrong, a number so low that it should not have any connection to law enforcement or justice. And in London, the Metropolitan police had been using facial recognition software that mistakenly identified an alarming 98% of people as criminals 1, which perhaps is a crime itself.

          MegaPixels is an online art project that explores the history of facial recognition from the perspective of datasets. To paraphrase the artist Trevor Paglen, whoever controls the dataset controls the meaning. MegaPixels aims to unravel the meanings behind the data and expose the darker corners of the biometric industry that have contributed to its growth. MegaPixels does not start with a conclusion, a moralistic slant, or a

          diff --git a/site/public/research/01_from_1_to_100_pixels/index.html b/site/public/research/01_from_1_to_100_pixels/index.html index 4446e1be..c11e966e 100644 --- a/site/public/research/01_from_1_to_100_pixels/index.html +++ b/site/public/research/01_from_1_to_100_pixels/index.html @@ -68,6 +68,9 @@
        1. NIST report on sres states several resolutions
        2. "Results show that the tested face recognition systems yielded similar performance for query sets with eye-to-eye distance from 60 pixels to 30 pixels" 1
          3. +
            +
          • "Note that we only keep the images with a minimal side length of 80 pixels." and "a face will be labeled as “Ignore” if it is very difficult to be detected due to blurring, severe deformation and unrecognizable eyes, or the side length of its bounding box is less than 32 pixels." Ge_Detecting_Masked_Faces_CVPR_2017_paper.pdf
            • +

          1. NIST 906932. Performance Assessment of Face Recognition Using Super-Resolution. Shuowen Hu, Robert Maschal, S. Susan Young, Tsai Hong Hong, Jonathon P. Phillips↩

            2. -- cgit v1.2.3-70-g09d2 From d8ea57ede73087e0590bc98c7a018f3f185d057a Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Thu, 28 Feb 2019 16:40:21 +0100 Subject: div over leaflet map --- client/map/index.js | 38 ++++++++++++++++++++++++++------------ site/assets/css/applets.css | 22 ++++++++++++++++++++++ 2 files changed, 48 insertions(+), 12 deletions(-) (limited to 'site') diff --git a/client/map/index.js b/client/map/index.js index d38855bf..53d9439d 100644 --- a/client/map/index.js +++ b/client/map/index.js @@ -2,25 +2,25 @@ import L from 'leaflet' import './leaflet.bezier' const arcStyles = { - 'edu': { + edu: { color: 'rgb(245, 246, 150)', fillColor: 'rgb(245, 246, 150)', opacity: 0.5, weight: '1', }, - 'company': { + company: { color: 'rgb(50, 100, 246)', fillColor: 'rgb(50, 100, 246)', opacity: 1.0, weight: '2', }, - 'gov': { + gov: { color: 'rgb(245, 150, 100)', fillColor: 'rgb(245, 150, 150)', opacity: 1.0, weight: '2', }, - 'mil': { + mil: { color: 'rgb(245, 0, 0)', fillColor: 'rgb(245, 0, 0)', opacity: 1.0, @@ -79,17 +79,31 @@ export default function append(el, payload) { } // ....i dont think the sort order does anything?? - citations.sort((a,b) => sortOrder.indexOf(a) - sortOrder.indexOf(b)) - .forEach(citation => { - const address = citation.addresses[0] - const latlng = [address.lat, address.lng].map(n => parseFloat(n)) - if (Number.isNaN(latlng[0]) || Number.isNaN(latlng[1])) return - addMarker(map, latlng, citation.title, address.name) - addArc(map, source, latlng, arcStyles[address.type]) - }) + citations.sort((a, b) => sortOrder.indexOf(a) - sortOrder.indexOf(b)) + .forEach(citation => { + const citationAddress = citation.addresses[0] + const latlng = [citationAddress.lat, citationAddress.lng].map(n => parseFloat(n)) + if (Number.isNaN(latlng[0]) || Number.isNaN(latlng[1])) return + addMarker(map, latlng, citation.title, citationAddress.name) + addArc(map, source, latlng, arcStyles[citationAddress.type]) + }) console.log(paper) const rootMarker = addMarker(map, source, paper.title, paper.address) rootMarker.openPopup() + + // a transparent div to cover the map, so normal scroll events will not be eaten by leaflet + const mapCover = document.createElement("div") + mapCover.classList.add("map_cover") + mapCover.innerHTML = "
            Click here to explore the map
            " + mapCover.querySelector('div').addEventListener('click', () => { + el.removeChild(mapCover) + }) + function stopPropagation(e) { + e.stopPropagation() + } + mapCover.addEventListener('mousewheel', stopPropagation, true) + mapCover.addEventListener('DOMMouseScroll', stopPropagation, true) + el.appendChild(mapCover) } diff --git a/site/assets/css/applets.css b/site/assets/css/applets.css index 729737fe..e84fcfc2 100644 --- a/site/assets/css/applets.css +++ b/site/assets/css/applets.css @@ -140,6 +140,28 @@ .map { margin-bottom: 20px; } +.map_cover { + position: absolute; + top: 0; + left: 0; + width: 100%; + height: 100%; + cursor: pointer; + background: rgba(0,0,0,0.8); + z-index: 9998; /* site header is 9999 */ + display: flex; + justify-content: center; + align-items: center; + font-size: 36px; + transition: opacity 0.4s cubic-bezier(0,0,1,1); + opacity: 1; +} +.desktop .map_cover { + opacity: 0; +} +.desktop .map_cover:hover { + opacity: 1; +} /* tabulator */ -- cgit v1.2.3-70-g09d2 From 0801726d7a3fd18fb7c4d1ec92e3581699d95ccc Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Thu, 28 Feb 2019 16:58:11 +0100 Subject: fix lfw commerical use csv --- client/tables.js | 1 + .../datasets/lfw/assets/lfw_commercial_use.csv | 62 +++++++++++----------- 2 files changed, 32 insertions(+), 31 deletions(-) (limited to 'site') diff --git a/client/tables.js b/client/tables.js index 2f4214e1..3fadb797 100644 --- a/client/tables.js +++ b/client/tables.js @@ -71,6 +71,7 @@ export default function append(el, payload) { table.setData(data) el.classList.add('loaded') } catch (e) { + console.error("error making json:", payload.url) console.error(e) // console.log(text) diff --git a/site/content/pages/datasets/lfw/assets/lfw_commercial_use.csv b/site/content/pages/datasets/lfw/assets/lfw_commercial_use.csv index 70e2fdeb..a2a4b39c 100644 --- a/site/content/pages/datasets/lfw/assets/lfw_commercial_use.csv +++ b/site/content/pages/datasets/lfw/assets/lfw_commercial_use.csv @@ -1,44 +1,44 @@ "name_display","company_url","example_url","country","description" -"Aratek","http://www.aratek.co/","","China","Biometric sensors for telecom, civil identification, finance, education, POS, and transportation" -"Asaphus","https://asaphus.de/","","Germany","Face recognition for home appliances and autonomous vehicles interaction" -"Aureus","https://cyberextruder.com/biometric-face-recognition-software-use-cases/","","USA","Retail loss prevention solutions, biometric access control, law enforcement and safe city applications, gaming and hospitality applications" +"Aratek","http://www.aratek.co/"," ","China","Biometric sensors for telecom, civil identification, finance, education, POS, and transportation" +"Asaphus","https://asaphus.de/"," ","Germany","Face recognition for home appliances and autonomous vehicles interaction" +"Aureus","https://cyberextruder.com/biometric-face-recognition-software-use-cases/"," ","USA","Retail loss prevention solutions, biometric access control, law enforcement and safe city applications, gaming and hospitality applications" "Baidu","http://research.baidu.com/institute-of-deep-learning/","https://www.newscientist.com/article/2113176-chinese-tourist-town-uses-face-recognition-as-an-entry-pass/","China","Retail payment, transportation, civil identification" -"Betaface","https://www.betaface.com/","","Germany","Web advertising and entertainment, video surveillance, security software, b2b software" -"Yi+AI","http://www.dress-plus.com/solution","","China","Scenario-based advertising, real-time personalized recommendation, character recognition for ads placement" -"CM-CV&AR","http://www.cloudminds.com/","","USA","Human augmented robot intelligence" -"Samtech","http://samtechinfonet.com/products_frs.php","","India","Facilities management, infrastructure support" -"ColorReco","http://www.colorreco.com/","","China","Face login verification, online payment security verification, access control system identity authentication and face recognition lock, mobile payment, driver fatigue recognition, virtual makeup" +"Betaface","https://www.betaface.com/"," ","Germany","Web advertising and entertainment, video surveillance, security software, b2b software" +"Yi+AI","http://www.dress-plus.com/solution"," ","China","Scenario-based advertising, real-time personalized recommendation, character recognition for ads placement" +"CM-CV&AR","http://www.cloudminds.com/"," ","USA","Human augmented robot intelligence" +"Samtech","http://samtechinfonet.com/products_frs.php"," ","India","Facilities management, infrastructure support" +"ColorReco","http://www.colorreco.com/"," ","China","Face login verification, online payment security verification, access control system identity authentication and face recognition lock, mobile payment, driver fatigue recognition, virtual makeup" "CloudWalk","www.cloudwalk.cn/","https://qz.com/africa/1287675/china-is-exporting-facial-recognition-to-africa-ensuring-ai-dominance-through-diversity/","China","Security and law enforcement. Being deployed in Zimbabwe" -"Cylltech","http://www.cylltech.com.cn/","","China","Conference management, social assistance, civil access, media orientation, precision marketing, scenic intelligence, tourism management" +"Cylltech","http://www.cylltech.com.cn/"," ","China","Conference management, social assistance, civil access, media orientation, precision marketing, scenic intelligence, tourism management" "Dahua-FaceImage","https://www.dahuasecurity.com/","https://www.dahuasecurity.com/solutions/solutionsbyapplication/23","China","Public security, public access control, finance" -"Daream","http://www.daream.com","","China","Fatigue and distraction detection for autonomous vehicles" -"Deepmark","https://deepmark.ru/","","Russia","Workplace access control" -"Easen Electron","http://www.easen-electron.com","","China","Face recognition door locks for automobiles" -"Ever AI","https://ever.ai/","","USA","Law enforcement, smart cities, surveillance, building security, retail, payments, autonomous vehicles, grocery stores, enhanced marketing" -"Facebook (Face.com)","https://en.wikipedia.org/wiki/Face.com","","USA","Sold to facebook in 2012, and now incorporated into DeepFace" -"Face++","https://www.faceplusplus.com/","","China","Audience engagement analysis, interactive marketing, gaming, photo album processing, security for mobile payments" -"Faceall","http://www.faceall.cn/index.en.html","","China","Internet banking, insurance, automated surveillance, access control, photo refinement, avatar creation" -"Faceter","https://faceter.io","","USA","Workforce attendence reporting and analytics, home video surveillance, retail customer behavior, GPU mining compatible" -"Facevisa","http://www.facevisa.com","","China","Face detection, face key point positioning, living body certification, facial attribute analysis" -"Fujitsu R&D","https://www.fujitsu.com/cn/en/about/local/subsidiaries/frdc/","","Japan","Consumer cameras" -"SenseTime","https://www.sensetime.com/","","Hong Kong","Surveillance, access control, image retrieval, and automatic log-on for personal computer or mobile devices" +"Daream","http://www.daream.com"," ","China","Fatigue and distraction detection for autonomous vehicles" +"Deepmark","https://deepmark.ru/"," ","Russia","Workplace access control" +"Easen Electron","http://www.easen-electron.com"," ","China","Face recognition door locks for automobiles" +"Ever AI","https://ever.ai/"," ","USA","Law enforcement, smart cities, surveillance, building security, retail, payments, autonomous vehicles, grocery stores, enhanced marketing" +"Facebook (Face.com)","https://en.wikipedia.org/wiki/Face.com"," ","USA","Sold to facebook in 2012, and now incorporated into DeepFace" +"Face++","https://www.faceplusplus.com/"," ","China","Audience engagement analysis, interactive marketing, gaming, photo album processing, security for mobile payments" +"Faceall","http://www.faceall.cn/index.en.html"," ","China","Internet banking, insurance, automated surveillance, access control, photo refinement, avatar creation" +"Faceter","https://faceter.io"," ","USA","Workforce attendence reporting and analytics, home video surveillance, retail customer behavior, GPU mining compatible" +"Facevisa","http://www.facevisa.com"," ","China","Face detection, face key point positioning, living body certification, facial attribute analysis" +"Fujitsu R&D","https://www.fujitsu.com/cn/en/about/local/subsidiaries/frdc/"," ","Japan","Consumer cameras" +"SenseTime","https://www.sensetime.com/"," ","Hong Kong","Surveillance, access control, image retrieval, and automatic log-on for personal computer or mobile devices" "Turing Robot","http://www.tuling123.com/","http://biz.turingos.cn/home","China","Emotion recognition and analysis for robots and toys, chatbots and digital assistants" "NEC","https://www.nec.com/en/press/201407/global_20140716_01.html","https://arxiv.org/abs/1212.6094","Japan","Law enforcement, event crowd monitoring, used specificallfy by Metropolitan police in UK" -"Aurora","http://auroracs.co.uk/","","UK","Face recognition in airports for security, queue management, x-ray divestment tray linkage" +"Aurora","http://auroracs.co.uk/"," ","UK","Face recognition in airports for security, queue management, x-ray divestment tray linkage" "VisionLabs","https://visionlabs.ai/","https://venturebeat.com/2016/07/07/russian-facial-recognition-startup-visionlabs-raises-5-5m-after-partnering-with-facebook-and-google/","Russia","Video surveillance, banking and finance, customer authentication for retail" -"Yunshitu","http://yunshitu.cn","","China","Security, Internet, broadcasting and other industries" -"Glasssix","http://www.glasssix.com/","","China","School attendance, workforce monitoring" +"Yunshitu","http://yunshitu.cn"," ","China","Security, Internet, broadcasting and other industries" +"Glasssix","http://www.glasssix.com/"," ","China","School attendance, workforce monitoring" "Hisign","http://www.hisign.com.cn/en-us/index.aspx","https://www.bloomberg.com/research/stocks/private/snapshot.asp?privcapId=52323181","China","Criminal investigation information application, and financial big data risk prevention and control products in China" "icarevision","http://www.icarevision.cn","https://www.bloomberg.com/research/stocks/private/snapshot.asp?privcapId=306707800","China","Video surveillance" "IntelliVision","https://www.intelli-vision.com/facial-recognition/","https://www.bloomberg.com/profiles/companies/0080393D:US-intellivision-technologies-corp","USA","Smart homes and buildings, smart security, smart city, smart retail, Smart auto" "Meiya Pico","https://meiyapico.com/","https://www.bloomberg.com/research/stocks/private/snapshot.asp?privcapId=117577345","China","Digital forensics and information security products and services in China" "Orion Star","https://www.ainirobot.com/#sixthPage","https://www.prnewswire.com/news-releases/orionstar-wins-challenge-to-recognize-one-million-celebrity-faces-with-artificial-intelligence-300494265.html","China","Face recognition for robots and livestream video censoring" -"Pegatron","http://www.pegatroncorp.com","","China","Workforce attendance" +"Pegatron","http://www.pegatroncorp.com"," ","China","Workforce attendance" "PingAn AI Lab","http://www.pingan.com/","https://www.biometricupdate.com/201703/ping-an-technology-developing-ai-face-recognition-technology-with-record-results","China","Financial services, lending" -"ReadSense","http://www.readsense.ai/","","China","Access control, traffic analysis, crowd analysis, head counting, drone vision, home appliances, community surveillance, custom attention analysis" -"sensingtech","www.sensingtech.com.cn","","China","Workplace entrypoint authentication" -"TCIT","http://www.tcit-us.com/?p=4023","","Taiwan","Retail analytics, workplace access control" -"TerminAI","terminai.com","","China","Smart office, smart city, smart gym, smart medical, smart community" -"Uni-Ubi","http://uni-ubi.com/","","China","Facial recognition for education, business, community, construction" -"Tencent YouTu Lab","http://bestimage.qq.com/","","China","Consumer applications for automatic facial beauty" -"Yuntu WiseSight","http://www.facelab.cn/","","China","Intrusion alarm, access control, access control, electronic patrol, and network alarm. detect suspicious personnel, real-name authentication, and public security, customs, airports, railways and other government security agencies, electronic patrol" \ No newline at end of file +"ReadSense","http://www.readsense.ai/"," ","China","Access control, traffic analysis, crowd analysis, head counting, drone vision, home appliances, community surveillance, custom attention analysis" +"sensingtech","www.sensingtech.com.cn"," ","China","Workplace entrypoint authentication" +"TCIT","http://www.tcit-us.com/?p=4023"," ","Taiwan","Retail analytics, workplace access control" +"TerminAI","terminai.com"," ","China","Smart office, smart city, smart gym, smart medical, smart community" +"Uni-Ubi","http://uni-ubi.com/"," ","China","Facial recognition for education, business, community, construction" +"Tencent YouTu Lab","http://bestimage.qq.com/"," ","China","Consumer applications for automatic facial beauty" +"Yuntu WiseSight","http://www.facelab.cn/"," ","China","Intrusion alarm, access control, access control, electronic patrol, and network alarm. detect suspicious personnel, real-name authentication, and public security, customs, airports, railways and other government security agencies, electronic patrol" \ No newline at end of file -- cgit v1.2.3-70-g09d2 From c33406d4da0f03a986db62b0d6b75c5a70114abe Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Thu, 28 Feb 2019 17:29:16 +0100 Subject: sidebar on about pages --- megapixels/app/site/parser.py | 12 +++++++----- site/assets/css/css.css | 4 ++++ site/content/pages/about/credits.md | 11 +++++++++++ site/content/pages/about/disclaimer.md | 11 +++++++++++ site/content/pages/about/index.md | 24 +++++++++++++++++++----- site/content/pages/about/press.md | 11 +++++++++++ site/content/pages/about/privacy.md | 10 ++++++++++ site/content/pages/about/terms.md | 12 +++++++++++- site/public/about/credits/index.html | 10 +++++++++- site/public/about/disclaimer/index.html | 10 +++++++++- site/public/about/index.html | 15 ++++++++++++--- site/public/about/press/index.html | 10 +++++++++- site/public/about/privacy/index.html | 11 +++++++++-- site/public/about/terms/index.html | 12 ++++++++++-- 14 files changed, 142 insertions(+), 21 deletions(-) (limited to 'site') diff --git a/megapixels/app/site/parser.py b/megapixels/app/site/parser.py index 9e904e00..b8bbf289 100644 --- a/megapixels/app/site/parser.py +++ b/megapixels/app/site/parser.py @@ -43,17 +43,19 @@ def parse_markdown(metadata, sections, s3_path, skip_h1=False): footnotes.append(section) elif ignoring: continue - elif '### Statistics' in section: + elif '### statistics' in section.lower() or '### sidebar' in section.lower(): if len(current_group): groups.append(format_section(current_group, s3_path)) current_group = [] - current_group.append(section) + if 'sidebar' not in section.lower(): + current_group.append(section) in_stats = True - elif in_stats and not section.strip().startswith('## '): + elif in_stats and not section.strip().startswith('## ') and 'end sidebar' not in section.lower(): current_group.append(section) - elif in_stats and section.strip().startswith('## '): + elif in_stats and section.strip().startswith('## ') or 'end sidebar' in section.lower(): current_group = [format_section(current_group, s3_path, 'right-sidebar', tag='div')] - current_group.append(section) + if 'end sidebar' not in section.lower(): + current_group.append(section) in_stats = False elif section.strip().startswith('```'): groups.append(format_section(current_group, s3_path)) diff --git a/site/assets/css/css.css b/site/assets/css/css.css index d710b3a8..ee99e13e 100644 --- a/site/assets/css/css.css +++ b/site/assets/css/css.css @@ -265,6 +265,10 @@ p { margin-bottom: 10px; color: #aaa; } +.right-sidebar ul:first-child a { + text-decoration: none; + border-bottom: 1px solid; +} /* lists */ diff --git a/site/content/pages/about/credits.md b/site/content/pages/about/credits.md index 2d16155c..3cd0b05b 100644 --- a/site/content/pages/about/credits.md +++ b/site/content/pages/about/credits.md @@ -12,6 +12,17 @@ authors: Adam Harvey # Credits +### Sidebar + +- [About](/about/) +- [Press](/about/press/) +- [Credits](/about/credits/) +- [Disclaimer](/about/disclaimer/) +- [Terms and Conditions](/about/terms/) +- [Privacy Policy](/about/privacy/) + +## End Sidebar + - MegaPixels by Adam Harvey - Made with support from Mozilla - Site developed by Jules Laplace diff --git a/site/content/pages/about/disclaimer.md b/site/content/pages/about/disclaimer.md index 64ce9f21..27cf6760 100644 --- a/site/content/pages/about/disclaimer.md +++ b/site/content/pages/about/disclaimer.md @@ -12,6 +12,17 @@ authors: Adam Harvey # Disclaimer +### Sidebar + +- [About](/about/) +- [Press](/about/press/) +- [Credits](/about/credits/) +- [Disclaimer](/about/disclaimer/) +- [Terms and Conditions](/about/terms/) +- [Privacy Policy](/about/privacy/) + +## End Sidebar + Last updated: December 04, 2018 The information contained on MegaPixels.cc website (the "Service") is for academic and artistic purposes only. diff --git a/site/content/pages/about/index.md b/site/content/pages/about/index.md index f9c6f83a..d3f5874d 100644 --- a/site/content/pages/about/index.md +++ b/site/content/pages/about/index.md @@ -12,18 +12,32 @@ authors: Adam Harvey # About MegaPixels -MegaPixels aims to answers to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla. +### Sidebar -MegaPixels aims to answers to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla. +- [Press](/about/press/) +- [Credits](/about/credits/) +- [Disclaimer](/about/disclaimer/) +- [Terms and Conditions](/about/terms/) +- [Privacy Policy](/about/privacy/) -+ Years: 2002-2004 ++ Years: 2002-2019 + Datasets Analyzed: 325 + Author: Adam Harvey + Development: Jules LaPlace + Research Assistance: Berit Gilma -![Adam Harvey](assets/adam-harvey.jpg) **Adam Harvey** is an American artist and researcher based in Berlin. His previous projects (CV Dazzle, Stealth Wear, and SkyLift) explore the potential for countersurveillance as artwork. He is the founder of VFRAME (visual forensics software for human rights groups), the recipient of 2 PrototypeFund awards, and is currently a researcher in residence at Karlsruhe HfG studying artifical intelligence and datasets. +## End Sidebar -![Adam Harvey](assets/jules-laplace.jpg) **Jules LaPlace** is an American technologist and artist also based in Berlin. He was previously the CTO for a NYC digital agency and currently works at VFRAME, developing computer vision for human rights groups, and as a freelance technologists for artists. +MegaPixels aims to answer to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla. + +MegaPixels aims to answer to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla. + +![Adam Harvey](assets/adam-harvey.jpg) + +**Adam Harvey** is an American artist and researcher based in Berlin. His previous projects (CV Dazzle, Stealth Wear, and SkyLift) explore the potential for countersurveillance as artwork. He is the founder of VFRAME (visual forensics software for human rights groups), the recipient of 2 PrototypeFund awards, and is currently a researcher in residence at Karlsruhe HfG studying artifical intelligence and datasets. + +![Jules LaPlace](assets/jules-laplace.jpg) + +**Jules LaPlace** is an American artist and technologist also based in Berlin. He was previously the CTO of a NYC digital agency and currently works at VFRAME, developing computer vision for human rights groups, and building creative software for artists. **Mozilla** is a free software community founded in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions. The community is supported institutionally by the not-for-profit Mozilla Foundation and its tax-paying subsidiary, the Mozilla Corporation. \ No newline at end of file diff --git a/site/content/pages/about/press.md b/site/content/pages/about/press.md index 2e3fa9a7..0e3124d0 100644 --- a/site/content/pages/about/press.md +++ b/site/content/pages/about/press.md @@ -13,6 +13,17 @@ authors: Adam Harvey # Press +### Sidebar + +- [About](/about/) +- [Press](/about/press/) +- [Credits](/about/credits/) +- [Disclaimer](/about/disclaimer/) +- [Terms and Conditions](/about/terms/) +- [Privacy Policy](/about/privacy/) + +## End Sidebar + ![alt text](assets/test.jpg) - Aug 22, 2018: "Transgender YouTubers had their videos grabbed to train facial recognition software" by James Vincent diff --git a/site/content/pages/about/privacy.md b/site/content/pages/about/privacy.md index 17d1b707..9685a189 100644 --- a/site/content/pages/about/privacy.md +++ b/site/content/pages/about/privacy.md @@ -12,6 +12,16 @@ authors: Adam Harvey # Privacy Policy +### Sidebar + +- [About](/about/) +- [Press](/about/press/) +- [Credits](/about/credits/) +- [Disclaimer](/about/disclaimer/) +- [Terms and Conditions](/about/terms/) +- [Privacy Policy](/about/privacy/) + +## End Sidebar A summary of our privacy policy is as follows: diff --git a/site/content/pages/about/terms.md b/site/content/pages/about/terms.md index 3735ff08..6ad03bc1 100644 --- a/site/content/pages/about/terms.md +++ b/site/content/pages/about/terms.md @@ -11,8 +11,18 @@ authors: Adam Harvey ------------ -Terms and Conditions ("Terms") +# Terms and Conditions ("Terms") +### Sidebar + +- [About](/about/) +- [Press](/about/press/) +- [Credits](/about/credits/) +- [Disclaimer](/about/disclaimer/) +- [Terms and Conditions](/about/terms/) +- [Privacy Policy](/about/privacy/) + +## End Sidebar Last updated: December 04, 2018 diff --git a/site/public/about/credits/index.html b/site/public/about/credits/index.html index fecc6c7b..6e4f06c1 100644 --- a/site/public/about/credits/index.html +++ b/site/public/about/credits/index.html @@ -28,7 +28,15 @@

            Credits

            -
              +
            +
            • MegaPixels by Adam Harvey
            • Made with support from Mozilla
            • Site developed by Jules Laplace
              • diff --git a/site/public/about/disclaimer/index.html b/site/public/about/disclaimer/index.html index a108baa0..b93194fa 100644 --- a/site/public/about/disclaimer/index.html +++ b/site/public/about/disclaimer/index.html @@ -28,7 +28,15 @@

              Disclaimer

              -

              Last updated: December 04, 2018

              +
              +

              Last updated: December 04, 2018

              The information contained on MegaPixels.cc website (the "Service") is for academic and artistic purposes only.

              MegaPixels.cc assumes no responsibility for errors or omissions in the contents on the Service.

              In no event shall MegaPixels.cc be liable for any special, direct, indirect, consequential, or incidental damages or any damages whatsoever, whether in an action of contract, negligence or other tort, arising out of or in connection with the use of the Service or the contents of the Service. MegaPixels.cc reserves the right to make additions, deletions, or modification to the contents on the Service at any time without prior notice.

              diff --git a/site/public/about/index.html b/site/public/about/index.html index 4a5ca926..2a0bc6c3 100644 --- a/site/public/about/index.html +++ b/site/public/about/index.html @@ -28,9 +28,18 @@

              About MegaPixels

              -

              MegaPixels aims to answers to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla.

              -

              MegaPixels aims to answers to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla.

              -
              Years
              2002-2004
              Datasets Analyzed
              325
              Author
              Adam Harvey
              Development
              Jules LaPlace
              Research Assistance
              Berit Gilma
              Adam Harvey
              Adam Harvey
              Adam Harvey
              Adam Harvey

              Mozilla is a free software community founded in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions. The community is supported institutionally by the not-for-profit Mozilla Foundation and its tax-paying subsidiary, the Mozilla Corporation.

              +
              +
              Years
              2002-2019
              Datasets Analyzed
              325
              Author
              Adam Harvey
              Development
              Jules LaPlace
              Research Assistance
              Berit Gilma

              MegaPixels aims to answer to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla.

              +

              MegaPixels aims to answer to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla.

              +
              Adam Harvey
              Adam Harvey

              Adam Harvey is an American artist and researcher based in Berlin. His previous projects (CV Dazzle, Stealth Wear, and SkyLift) explore the potential for countersurveillance as artwork. He is the founder of VFRAME (visual forensics software for human rights groups), the recipient of 2 PrototypeFund awards, and is currently a researcher in residence at Karlsruhe HfG studying artifical intelligence and datasets.

              +
              Jules LaPlace
              Jules LaPlace

              Jules LaPlace is an American artist and technologist also based in Berlin. He was previously the CTO of a NYC digital agency and currently works at VFRAME, developing computer vision for human rights groups, and building creative software for artists.

              +

              Mozilla is a free software community founded in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions. The community is supported institutionally by the not-for-profit Mozilla Foundation and its tax-paying subsidiary, the Mozilla Corporation.

              diff --git a/site/public/about/press/index.html b/site/public/about/press/index.html index a1d9d4f5..d36b6bc6 100644 --- a/site/public/about/press/index.html +++ b/site/public/about/press/index.html @@ -28,7 +28,15 @@

              Press

              -
              alt text
              alt text
                +
              +
              alt text
              alt text
              • Aug 22, 2018: "Transgender YouTubers had their videos grabbed to train facial recognition software" by James Vincent https://www.theverge.com/2017/8/22/16180080/transgender-youtubers-ai-facial-recognition-dataset
              • Aug 22, 2018: "Transgender YouTubers had their videos grabbed to train facial recognition software" by James Vincent https://www.theverge.com/2017/8/22/16180080/transgender-youtubers-ai-facial-recognition-dataset
              • Aug 22, 2018: "Transgender YouTubers had their videos grabbed to train facial recognition software" by James Vincent https://www.theverge.com/2017/8/22/16180080/transgender-youtubers-ai-facial-recognition-dataset diff --git a/site/public/about/privacy/index.html b/site/public/about/privacy/index.html index 92a1b9a8..1b3b9d2f 100644 --- a/site/public/about/privacy/index.html +++ b/site/public/about/privacy/index.html @@ -28,10 +28,17 @@

                Privacy Policy

                -

                A summary of our privacy policy is as follows:

                +
                +

                A summary of our privacy policy is as follows:

                The MegaPixels site does not use any analytics programs or collect any data besides the necessary IP address of your connection, which are deleted every 30 days and used only for security and to prevent misuse.

                The image processing sections of the site do not collect any data whatsoever. All processing takes place in temporary memory (RAM) and then is displayed back to the user over a SSL secured HTTPS connection. It is the sole responsibility of the user whether they discard, by closing the page, or share their analyzed information and any potential consequences that may arise from doing so.

                -

                A more complete legal version is below:

                This is a boilerplate Privacy policy from https://termsfeed.com/

                Needs to be reviewed

                diff --git a/site/public/about/terms/index.html b/site/public/about/terms/index.html index fd17b4d9..8bd6e738 100644 --- a/site/public/about/terms/index.html +++ b/site/public/about/terms/index.html @@ -27,8 +27,16 @@
                -

                Terms and Conditions ("Terms")

                -

                Last updated: December 04, 2018

                +

                Terms and Conditions ("Terms")

                +
                +

                Last updated: December 04, 2018

                Please read these Terms and Conditions ("Terms", "Terms and Conditions") carefully before using the MegaPixels website (the "Service") operated by megapixels.cc ("us", "we", or "our").

                Your access to and use of the Service is conditioned on your acceptance of and compliance with these Terms.

                By accessing or using the Service you agree to be bound by these Terms. If you disagree with any part of the terms then you may not access the Service.

                -- cgit v1.2.3-70-g09d2 From 18e595bdaf64417622d12fcbe9b5af96ac935ab3 Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Thu, 28 Feb 2019 17:39:22 +0100 Subject: special case adam/jules sideimages --- megapixels/app/site/parser.py | 9 ++++++--- site/assets/css/css.css | 11 ++++++++++- site/content/pages/about/index.md | 8 +++----- site/public/about/index.html | 6 +++--- 4 files changed, 22 insertions(+), 12 deletions(-) (limited to 'site') diff --git a/megapixels/app/site/parser.py b/megapixels/app/site/parser.py index b8bbf289..c17d3b8a 100644 --- a/megapixels/app/site/parser.py +++ b/megapixels/app/site/parser.py @@ -144,7 +144,7 @@ def intro_section(metadata, s3_path): def fix_images(lines, s3_path): """ - do our own tranformation of the markdown around images to handle wide images etc + do our own transformation of the markdown around images to handle wide images etc lines: markdown lines """ real_lines = [] @@ -154,10 +154,13 @@ def fix_images(lines, s3_path): line = line.replace('![', '') alt_text, tail = line.split('](', 1) url, tail = tail.split(')', 1) + tag = '' if ':' in alt_text: - tail, alt_text = alt_text.split(':', 1) + tag, alt_text = alt_text.split(':', 1) img_tag = "{}".format(s3_path + url, alt_text.replace("'", "")) - if len(alt_text): + if 'sideimage' in tag: + line = "
                {}
                {}
                ".format(img_tag, markdown(tail)) + elif len(alt_text): line = "
                {}
                {}
                ".format(img_tag, alt_text) else: line = "
                {}
                ".format(img_tag, alt_text) diff --git a/site/assets/css/css.css b/site/assets/css/css.css index ee99e13e..29833be7 100644 --- a/site/assets/css/css.css +++ b/site/assets/css/css.css @@ -387,7 +387,16 @@ section.fullwidth .image { margin: 10px auto 0 auto; font-family: 'Roboto'; } - +.sideimage { + margin: 10px 0; + display: flex; + flex-direction: row; + justify-content: flex-start; + align-items: flex-start; +} +.sideimage img { + margin-right: 10px; +} /* blog index */ .research_index { diff --git a/site/content/pages/about/index.md b/site/content/pages/about/index.md index d3f5874d..861cfd07 100644 --- a/site/content/pages/about/index.md +++ b/site/content/pages/about/index.md @@ -32,12 +32,10 @@ MegaPixels aims to answer to these questions and reveal the stories behind the m MegaPixels aims to answer to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla. -![Adam Harvey](assets/adam-harvey.jpg) +![sideimage:Adam Harvey](assets/adam-harvey.jpg) **Adam Harvey** is an American artist and researcher based in Berlin. His previous projects (CV Dazzle, Stealth Wear, and SkyLift) explore the potential for countersurveillance as artwork. He is the founder of VFRAME (visual forensics software for human rights groups), the recipient of 2 PrototypeFund awards, and is currently a researcher in residence at Karlsruhe HfG studying artifical intelligence and datasets. -**Adam Harvey** is an American artist and researcher based in Berlin. His previous projects (CV Dazzle, Stealth Wear, and SkyLift) explore the potential for countersurveillance as artwork. He is the founder of VFRAME (visual forensics software for human rights groups), the recipient of 2 PrototypeFund awards, and is currently a researcher in residence at Karlsruhe HfG studying artifical intelligence and datasets. +![sideimage:Jules LaPlace](assets/jules-laplace.jpg) **Jules LaPlace** is an American artist and technologist also based in Berlin. He was previously the CTO of a NYC digital agency and currently works at VFRAME, developing computer vision for human rights groups, and building creative software for artists. -![Jules LaPlace](assets/jules-laplace.jpg) +**Mozilla** is a free software community founded in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions. The community is supported institutionally by the not-for-profit Mozilla Foundation and its tax-paying subsidiary, the Mozilla Corporation. -**Jules LaPlace** is an American artist and technologist also based in Berlin. He was previously the CTO of a NYC digital agency and currently works at VFRAME, developing computer vision for human rights groups, and building creative software for artists. -**Mozilla** is a free software community founded in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions. The community is supported institutionally by the not-for-profit Mozilla Foundation and its tax-paying subsidiary, the Mozilla Corporation. \ No newline at end of file diff --git a/site/public/about/index.html b/site/public/about/index.html index 2a0bc6c3..8583fd96 100644 --- a/site/public/about/index.html +++ b/site/public/about/index.html @@ -37,9 +37,9 @@
              Years
              2002-2019
              Datasets Analyzed
              325
              Author
              Adam Harvey
              Development
              Jules LaPlace
              Research Assistance
              Berit Gilma

              MegaPixels aims to answer to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla.

              MegaPixels aims to answer to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla.

              -
              Adam Harvey
              Adam Harvey

              Adam Harvey is an American artist and researcher based in Berlin. His previous projects (CV Dazzle, Stealth Wear, and SkyLift) explore the potential for countersurveillance as artwork. He is the founder of VFRAME (visual forensics software for human rights groups), the recipient of 2 PrototypeFund awards, and is currently a researcher in residence at Karlsruhe HfG studying artifical intelligence and datasets.

              -
              Jules LaPlace
              Jules LaPlace

              Jules LaPlace is an American artist and technologist also based in Berlin. He was previously the CTO of a NYC digital agency and currently works at VFRAME, developing computer vision for human rights groups, and building creative software for artists.

              -

              Mozilla is a free software community founded in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions. The community is supported institutionally by the not-for-profit Mozilla Foundation and its tax-paying subsidiary, the Mozilla Corporation.

              +
              Adam Harvey

              Adam Harvey is an American artist and researcher based in Berlin. His previous projects (CV Dazzle, Stealth Wear, and SkyLift) explore the potential for countersurveillance as artwork. He is the founder of VFRAME (visual forensics software for human rights groups), the recipient of 2 PrototypeFund awards, and is currently a researcher in residence at Karlsruhe HfG studying artifical intelligence and datasets.

              +
              Jules LaPlace

              Jules LaPlace is an American artist and technologist also based in Berlin. He was previously the CTO of a NYC digital agency and currently works at VFRAME, developing computer vision for human rights groups, and building creative software for artists.

              +

              Mozilla is a free software community founded in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions. The community is supported institutionally by the not-for-profit Mozilla Foundation and its tax-paying subsidiary, the Mozilla Corporation.

              -- cgit v1.2.3-70-g09d2 From ef90adeb4230ac27c18d3ed9e2cfab000c8689e0 Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Thu, 28 Feb 2019 18:09:27 +0100 Subject: recreate dataset index --- megapixels/app/site/builder.py | 2 +- site/assets/css/css.css | 13 +++- site/content/pages/about/index.md | 2 - site/content/pages/datasets/index.md | 27 ++++++++ site/content/pages/datasets/lfw/index.md | 2 +- site/public/datasets/index.html | 71 +++++++++++++++++++--- site/public/datasets/lfw/index.html | 2 +- site/public/datasets/vgg_face2/index.html | 2 +- site/public/datasets_v0/index.html | 2 +- site/public/datasets_v0/lfw/index.html | 2 +- .../datasets_v0/lfw/right-to-removal/index.html | 1 - site/public/datasets_v0/vgg_face2/index.html | 2 +- site/public/index.html | 6 ++ site/templates/datasets.html | 9 ++- 14 files changed, 116 insertions(+), 27 deletions(-) create mode 100644 site/content/pages/datasets/index.md (limited to 'site') diff --git a/megapixels/app/site/builder.py b/megapixels/app/site/builder.py index 15055110..603d4788 100644 --- a/megapixels/app/site/builder.py +++ b/megapixels/app/site/builder.py @@ -78,7 +78,7 @@ def build_index(key, research_posts, datasets): template = env.get_template("page.html") s3_path = s3.make_s3_path(cfg.S3_SITE_PATH, metadata['path']) content = parser.parse_markdown(metadata, sections, s3_path, skip_h1=False) - content += loader.parse_research_index(research_posts) + content += parser.parse_research_index(research_posts) html = template.render( metadata=metadata, content=content, diff --git a/site/assets/css/css.css b/site/assets/css/css.css index 29833be7..3bd09f23 100644 --- a/site/assets/css/css.css +++ b/site/assets/css/css.css @@ -1,4 +1,4 @@ -* { box-sizing: border-box; outline: 0; } +da* { box-sizing: border-box; outline: 0; } html, body { margin: 0; padding: 0; @@ -396,7 +396,10 @@ section.fullwidth .image { } .sideimage img { margin-right: 10px; + width: 250px; + height: 250px; } + /* blog index */ .research_index { @@ -521,7 +524,8 @@ section.fullwidth .image { text-decoration: none; transition: background-color 0.1s cubic-bezier(0,0,1,1); background: black; - margin: 0 20px 20px 0; + margin: 0 11px 11px 0; + border: 0; } .dataset-list .dataset { width: 220px; @@ -538,6 +542,11 @@ section.fullwidth .image { .dataset-list a:nth-child(3n+3) { background-color: rgba(255, 255, 0, 0.1); } .desktop .dataset-list .dataset:nth-child(3n+3):hover { background-color: rgba(255, 255, 0, 0.2); } +.dataset-list span { + box-shadow: -3px -3px black, 3px -3px black, -3px 3px black, 3px 3px black; + background-color: black; + box-decoration-break: clone; +} /* intro section for datasets */ diff --git a/site/content/pages/about/index.md b/site/content/pages/about/index.md index 861cfd07..66fac8ae 100644 --- a/site/content/pages/about/index.md +++ b/site/content/pages/about/index.md @@ -37,5 +37,3 @@ MegaPixels aims to answer to these questions and reveal the stories behind the m ![sideimage:Jules LaPlace](assets/jules-laplace.jpg) **Jules LaPlace** is an American artist and technologist also based in Berlin. He was previously the CTO of a NYC digital agency and currently works at VFRAME, developing computer vision for human rights groups, and building creative software for artists. **Mozilla** is a free software community founded in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions. The community is supported institutionally by the not-for-profit Mozilla Foundation and its tax-paying subsidiary, the Mozilla Corporation. - - diff --git a/site/content/pages/datasets/index.md b/site/content/pages/datasets/index.md new file mode 100644 index 00000000..c408fba4 --- /dev/null +++ b/site/content/pages/datasets/index.md @@ -0,0 +1,27 @@ +------------ + +status: published +title: MegaPixels: Datasets +desc: Facial Recognition Datasets +slug: home +published: 2018-12-15 +updated: 2018-12-15 +authors: Adam Harvey +sync: false + +------------ + +# Facial Recognition Datasets + +### Sidebar + ++ Found: 275 datasets ++ Created between: 1993-2018 ++ Smallest dataset: 20 images ++ Largest dataset: 10,000,000 images + ++ Highest resolution faces: 450x500 (Unconstrained College Students) ++ Lowest resolution faces: 16x20 pixels (QMUL SurvFace) + +## End Sidebar + diff --git a/site/content/pages/datasets/lfw/index.md b/site/content/pages/datasets/lfw/index.md index 972fafe2..4161561d 100644 --- a/site/content/pages/datasets/lfw/index.md +++ b/site/content/pages/datasets/lfw/index.md @@ -4,7 +4,7 @@ status: published title: Labeled Faces in The Wild desc: Labeled Faces in The Wild (LFW) is a database of face photographs designed for studying the problem of unconstrained face recognition. subdesc: It includes 13,456 images of 4,432 people’s images copied from the Internet during 2002-2004. -image: assets/lfw_feature.jpg +image: assets/background.jpg caption: A few of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms. slug: lfw published: 2019-2-23 diff --git a/site/public/datasets/index.html b/site/public/datasets/index.html index 77c5ab2b..17c938ac 100644 --- a/site/public/datasets/index.html +++ b/site/public/datasets/index.html @@ -29,27 +29,78 @@

              Facial Recognition Datasets

              -

              Regular Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.

              -

              Summary

              -
              Found
              275 datasets
              Created between
              1993-2018
              Smallest dataset
              20 images
              Largest dataset
              10,000,000 images
              Highest resolution faces
              450x500 (Unconstrained College Students)
              Lowest resolution faces
              16x20 pixels (QMUL SurvFace)
              +
            Found
            275 datasets
            Created between
            1993-2018
            Smallest dataset
            20 images
            Largest dataset
            10,000,000 images
            Highest resolution faces
            450x500 (Unconstrained College Students)
            Lowest resolution faces
            16x20 pixels (QMUL SurvFace)
            -
            -

            Dataset Portraits

            +

            - We have prepared detailed studies of some of the more noteworthy datasets. + We have prepared detailed case studies of some of the more noteworthy datasets, including tools to help you learn what is contained in these datasets, and even whether your own face has been used to train these algorithms.

            - +
            - Labeled Faces in The Wild + Asian Face Age Dataset
            - +
            - VGG Face2 + Annotated Facial Landmarks in The Wild +
            +             + + +
            + Caltech 10K Faces Dataset +
            +             + + +
            + Caltech Occluded Faces in The Wild +
            +             + + +
            + Facebook +
            +             + + +
            + FERET: FacE REcognition +
            +             + + +
            + Labeled Face Parts in The Wild +
            +             + + +
            + Labeled Faces in The Wild +
            +             + + +
            + Unconstrained College Students +
            +             + + +
            + VGG Face 2 Dataset +
            +             + + +
            + YouTube Celebrities
            diff --git a/site/public/datasets/lfw/index.html b/site/public/datasets/lfw/index.html index 08ec8ee3..5b5e58f3 100644 --- a/site/public/datasets/lfw/index.html +++ b/site/public/datasets/lfw/index.html @@ -27,7 +27,7 @@
            -
            Labeled Faces in The Wild (LFW) is a database of face photographs designed for studying the problem of unconstrained face recognition.
            It includes 13,456 images of 4,432 people’s images copied from the Internet during 2002-2004. +
            Labeled Faces in The Wild (LFW) is a database of face photographs designed for studying the problem of unconstrained face recognition.
            It includes 13,456 images of 4,432 people’s images copied from the Internet during 2002-2004.
            A few of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.

            Statistics

            Years
            2002-2004
            Images
            13,233
            Identities
            5,749
            Origin
            Yahoo News Images
            Funding
            (Possibly, partially CIA)

            INSIGHTS

              diff --git a/site/public/datasets/vgg_face2/index.html b/site/public/datasets/vgg_face2/index.html index 08b02cc7..efe6cb84 100644 --- a/site/public/datasets/vgg_face2/index.html +++ b/site/public/datasets/vgg_face2/index.html @@ -28,7 +28,7 @@

              VGG Face 2

              -
              Years
              TBD
              Images
              TBD
              Identities
              TBD
              Origin
              TBD
              Funding
              IARPA
              ...
              ...

              Analysis

              +
              Years
              TBD
              Images
              TBD
              Identities
              TBD
              Origin
              TBD
              Funding
              IARPA
              ...
              ...

              Analysis

              • The VGG Face 2 dataset includes approximately 1,331 actresses, 139 presidents, 16 wives, 3 husbands, 2 snooker player, and 1 guru
              diff --git a/site/public/datasets_v0/index.html b/site/public/datasets_v0/index.html index 71147a64..c2e6617b 100644 --- a/site/public/datasets_v0/index.html +++ b/site/public/datasets_v0/index.html @@ -30,7 +30,7 @@

              Facial Recognition Datasets

              Regular Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.

              Summary

              -
              Found
              275 datasets
              Created between
              1993-2018
              Smallest dataset
              20 images
              Largest dataset
              10,000,000 images
              Highest resolution faces
              450x500 (Unconstrained College Students)
              Lowest resolution faces
              16x20 pixels (QMUL SurvFace)
              +
              Found
              275 datasets
              Created between
              1993-2018
              Smallest dataset
              20 images
              Largest dataset
              10,000,000 images
              Highest resolution faces
              450x500 (Unconstrained College Students)
              Lowest resolution faces
              16x20 pixels (QMUL SurvFace)
              diff --git a/site/public/datasets_v0/lfw/index.html b/site/public/datasets_v0/lfw/index.html index b4ee82a3..4ee4799f 100644 --- a/site/public/datasets_v0/lfw/index.html +++ b/site/public/datasets_v0/lfw/index.html @@ -28,7 +28,7 @@

              Labeled Faces in the Wild

              -
              Created
              2007
              Images
              13,233
              People
              5,749
              Created From
              Yahoo News images
              Search available
              Searchable
              Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.
              Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.

              Intro

              +
              Created
              2007
              Images
              13,233
              People
              5,749
              Created From
              Yahoo News images
              Search available
              Searchable
              Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.
              Eighteen of the 5,749 people in the Labeled Faces in the Wild Dataset. The most widely used face dataset for benchmarking commercial face recognition algorithms.

              Intro

              Labeled Faces in The Wild (LFW) is among the most widely used facial recognition training datasets in the world and is the first of its kind to be created entirely from images posted online. The LFW dataset includes 13,233 images of 5,749 people that were collected between 2002-2004. Use the tools below to check if you were included in this dataset or scroll down to read the analysis.

              Three paragraphs describing the LFW dataset in a format that can be easily replicated for the other datasets. Nothing too custom. An analysis of the initial research papers with context relative to all the other dataset papers.

              From George W. Bush to Jamie Lee Curtis: all 5,749 people in the LFW Dataset sorted from most to least images collected.
              From George W. Bush to Jamie Lee Curtis: all 5,749 people in the LFW Dataset sorted from most to least images collected.

              LFW by the Numbers

              diff --git a/site/public/datasets_v0/lfw/right-to-removal/index.html b/site/public/datasets_v0/lfw/right-to-removal/index.html index 5dc269b2..97ce4d05 100644 --- a/site/public/datasets_v0/lfw/right-to-removal/index.html +++ b/site/public/datasets_v0/lfw/right-to-removal/index.html @@ -38,7 +38,6 @@

              The dataset is being used thousands of companies around the world to improve facial recognition software including usage by governments for the purpose of law enforcement, national security, tracking consumers in retail environments, and tracking individuals through public spaces.

              My name as it appears in your dataset is [your name]. Please remove all images from your dataset and inform your newsletter subscribers to likewise update their copies.

              - [your name]

              -
              diff --git a/site/public/datasets_v0/vgg_face2/index.html b/site/public/datasets_v0/vgg_face2/index.html index b7ba5a4c..6a67e7e4 100644 --- a/site/public/datasets_v0/vgg_face2/index.html +++ b/site/public/datasets_v0/vgg_face2/index.html @@ -28,7 +28,7 @@

              VGG Faces2

              -
              Created
              2018
              Images
              3.3M
              People
              9,000
              Created From
              Scraping search engines
              Search available
              [Searchable](#)

              VGG Face2 is the updated version of the VGG Face dataset and now includes over 3.3M face images from over 9K people. The identities were selected by taking the top 500K identities in Google's Knowledge Graph of celebrities and then selecting only the names that yielded enough training images. The dataset was created in the UK but funded by Office of Director of National Intelligence in the United States.

              +
              Created
              2018
              Images
              3.3M
              People
              9,000
              Created From
              Scraping search engines
              Search available
              [Searchable](#)

              VGG Face2 is the updated version of the VGG Face dataset and now includes over 3.3M face images from over 9K people. The identities were selected by taking the top 500K identities in Google's Knowledge Graph of celebrities and then selecting only the names that yielded enough training images. The dataset was created in the UK but funded by Office of Director of National Intelligence in the United States.

              VGG Face2 by the Numbers

              • 1,331 actresses, 139 presidents
                • diff --git a/site/public/index.html b/site/public/index.html index 5038c483..8775f22d 100644 --- a/site/public/index.html +++ b/site/public/index.html @@ -110,6 +110,12 @@
              + +
              + Facebook +
              +               +
              FERET: FacE REcognition diff --git a/site/templates/datasets.html b/site/templates/datasets.html index ba230eee..3456eac8 100644 --- a/site/templates/datasets.html +++ b/site/templates/datasets.html @@ -4,17 +4,16 @@ {{ content }} -
              -

              Dataset Portraits

              +

              - We have prepared detailed studies of some of the more noteworthy datasets. + We have prepared detailed case studies of some of the more noteworthy datasets, including tools to help you learn what is contained in these datasets, and even whether your own face has been used to train these algorithms.

              {% for dataset in datasets %} - +
              - {{ dataset.title }} + {{ dataset.title }}
              {% endfor %} -- cgit v1.2.3-70-g09d2 From 406d857c61fb128a48281a52899ddf77b68201be Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Thu, 28 Feb 2019 18:32:39 +0100 Subject: threejs splash page on the index --- client/splash/index.js | 14 +-- megapixels/app/site/parser.py | 2 + site/assets/css/css.css | 15 +++- site/content/pages/datasets/index.md | 5 -- site/content/pages/index.md | 26 ++---- site/content/pages/info/index.md | 2 +- site/public/datasets/index.html | 2 +- site/public/index.html | 166 +++-------------------------------- site/public/info/index.html | 2 +- site/templates/home.html | 113 ++++++++---------------- 10 files changed, 74 insertions(+), 273 deletions(-) (limited to 'site') diff --git a/client/splash/index.js b/client/splash/index.js index e247b7f5..a21110f0 100644 --- a/client/splash/index.js +++ b/client/splash/index.js @@ -31,12 +31,14 @@ function build() { function bind() { document.querySelector('.slogan').addEventListener('click', modal.close) - toArray(document.querySelectorAll('.aboutLink')).forEach(el => { - el.addEventListener('click', modal.toggle) - }) - document.querySelector('.about .inner').addEventListener('click', e => e.stopPropagation()) - document.querySelector('.about').addEventListener('click', modal.close) - document.querySelector('.close').addEventListener('click', modal.close) + if (document.querySelector('.about')) { + toArray(document.querySelectorAll('.aboutLink')).forEach(el => { + el.addEventListener('click', modal.toggle) + }) + document.querySelector('.about .inner').addEventListener('click', e => e.stopPropagation()) + document.querySelector('.about').addEventListener('click', modal.close) + document.querySelector('.close').addEventListener('click', modal.close) + } } function animate() { diff --git a/megapixels/app/site/parser.py b/megapixels/app/site/parser.py index c17d3b8a..ad4256ad 100644 --- a/megapixels/app/site/parser.py +++ b/megapixels/app/site/parser.py @@ -198,6 +198,8 @@ def format_metadata(section): """ meta = [] for line in section.split('\n'): + if ': ' not in line: + continue key, value = line[2:].split(': ', 1) meta.append("
              {}
              {}
              ".format(key, value)) return "
              {}
              ".format(''.join(meta)) diff --git a/site/assets/css/css.css b/site/assets/css/css.css index 3bd09f23..732386bd 100644 --- a/site/assets/css/css.css +++ b/site/assets/css/css.css @@ -112,13 +112,19 @@ footer { justify-content: space-between; color: #888; font-size: 9pt; - padding: 20px 75px 20px; + padding: 20px 0 20px; font-family: "Roboto", sans-serif; } footer > div { display: flex; flex-direction: row; } +footer > div:nth-child(1) { + padding-left: 75px; +} +footer > div:nth-child(2) { + padding-right: 75px; +} footer a { display: inline-block; color: #888; @@ -237,6 +243,7 @@ p { align-items: flex-start; font-size: 10pt; margin-bottom: 20px; + font-family: 'Roboto', sans-serif; } .meta > div { margin-right: 30px; @@ -540,11 +547,11 @@ section.fullwidth .image { .desktop .dataset-list a:nth-child(3n+2):hover { background-color: rgba(255, 128, 0, 0.2); } .dataset-list a:nth-child(3n+3) { background-color: rgba(255, 255, 0, 0.1); } -.desktop .dataset-list .dataset:nth-child(3n+3):hover { background-color: rgba(255, 255, 0, 0.2); } +.desktop .dataset-list a:nth-child(3n+3):hover { background-color: rgba(255, 255, 0, 0.2); } .dataset-list span { - box-shadow: -3px -3px black, 3px -3px black, -3px 3px black, 3px 3px black; - background-color: black; + box-shadow: -3px -3px #181818, 3px -3px #181818, -3px 3px #181818, 3px 3px #181818; + background-color: #181818; box-decoration-break: clone; } diff --git a/site/content/pages/datasets/index.md b/site/content/pages/datasets/index.md index c408fba4..fa012758 100644 --- a/site/content/pages/datasets/index.md +++ b/site/content/pages/datasets/index.md @@ -13,8 +13,6 @@ sync: false # Facial Recognition Datasets -### Sidebar - + Found: 275 datasets + Created between: 1993-2018 + Smallest dataset: 20 images @@ -22,6 +20,3 @@ sync: false + Highest resolution faces: 450x500 (Unconstrained College Students) + Lowest resolution faces: 16x20 pixels (QMUL SurvFace) - -## End Sidebar - diff --git a/site/content/pages/index.md b/site/content/pages/index.md index d63cf9fa..1cf47aac 100644 --- a/site/content/pages/index.md +++ b/site/content/pages/index.md @@ -1,30 +1,14 @@ ------------ status: published -title: MegaPixels -desc: -slug: home +title: Megapixels +desc: The Darkside of Datasets +slug: analysis published: 2018-12-15 updated: 2018-12-15 authors: Adam Harvey sync: false - ------------- - -## Facial Recognition Datasets - -Regular Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. -### Summary - -+ Found: 275 datasets -+ Created between: 1993-2018 -+ Smallest dataset: 20 images -+ Largest dataset: 10,000,000 images - -+ Highest resolution faces: 450x500 (Unconstrained College Students) -+ Lowest resolution faces: 16x20 pixels (QMUL SurvFace) +------------ -``` -load_file https://megapixels.nyc3.digitaloceanspaces.com/v1/citations/datasets.csv -``` +## diff --git a/site/content/pages/info/index.md b/site/content/pages/info/index.md index 9cbb219e..090783d9 100644 --- a/site/content/pages/info/index.md +++ b/site/content/pages/info/index.md @@ -11,7 +11,7 @@ sync: false ------------ -## +## Face Analysis ``` face_analysis diff --git a/site/public/datasets/index.html b/site/public/datasets/index.html index 17c938ac..7398da17 100644 --- a/site/public/datasets/index.html +++ b/site/public/datasets/index.html @@ -29,7 +29,7 @@

              Facial Recognition Datasets

              -
              Found
              275 datasets
              Created between
              1993-2018
              Smallest dataset
              20 images
              Largest dataset
              10,000,000 images
              Highest resolution faces
              450x500 (Unconstrained College Students)
              Lowest resolution faces
              16x20 pixels (QMUL SurvFace)
              +
              Found
              275 datasets
              Created between
              1993-2018
              Smallest dataset
              20 images
              Largest dataset
              10,000,000 images
              Highest resolution faces
              450x500 (Unconstrained College Students)
              Lowest resolution faces
              16x20 pixels (QMUL SurvFace)

              diff --git a/site/public/index.html b/site/public/index.html index 8775f22d..d5a2e59f 100644 --- a/site/public/index.html +++ b/site/public/index.html @@ -3,15 +3,13 @@ MegaPixels - - + + - - - +

              @@ -20,166 +18,22 @@
              MegaPixels
              - Datasets - Research - About + DATASETS + RESEARCH + ABOUT
              -
              - -
              -
              -
              -
              -
              -
              -
              - MegaPixels is an art project that explores the dark side of face recognition datasets and the future of computer vision. -
              - -
              - - -
              - -
              - Made by Adam Harvey in collaboration with Jules Laplace, and in partnership with Mozilla.
              - Read more about MegaPixels -
              -
              -
              -
              - -
              -

              Face Recognition Datasets

              -
              -

              SUMMARY

              -
              -
              Found
              275 datasets
              -
              Created between
              1993-2018
              -
              Smallest dataset
              20 images
              -
              Largest dataset
              10,000,000 images
              -
              Highest resolution faces
              450x500 (Unconstrained College Students)
              -
              Lowest resolution faces
              16x20 pixels (QMUL SurvFace)
              -
              -
              - -

              - MegaPixels is an online art project that explores the history of face recognition from the perspective of datasets. MegaPixels aims to unravel the meanings behind the data and expose the darker corners of the biometric industry that have contributed to its growth. -

              -

              - Through a mix of case studies, visualizations, and interactive tools, Megapixels will use face recognition datasets to tell the history of modern biometrics. Many people have contributed to the development of face recignition technology, both wittingly and unwittingly. Not only scientists, but also celebrities and regular internet users have played a part. -

              -

              - Face recognition is a mess of contradictinos. It works, yet it doesn't actually work. It's cheap and accessible, but also expensive and out of control. Face recognition research has achieved headline grabbing superhuman accuracies over 99.9%, yet in practice it's also dangerously inaccurate. -

              -

              - During a trial installation at Sudkreuz station in Berlin in 2018, 20% of the matches were wrong, a number so low that it should not have any connection to law enforcement or justice. And in London, the Metropolitan police had been using face recognition software that mistakenly identified an alarming 98% of people as criminals, which perhaps is a crime itself. -

              -
              - -
              -

              Dataset Portraits

              -

              - We have prepared detailed case studies of some of the more noteworthy datasets, including tools to help you learn what is contained in these datasets, and even whether your own face has been used to train these algorithms. -

              - -
              - - -
              - Asian Face Age Dataset -
              -               - - -
              - Annotated Facial Landmarks in The Wild -
              -               - - -
              - Caltech 10K Faces Dataset -
              -               - - -
              - Caltech Occluded Faces in The Wild -
              -               - - -
              - Facebook -
              -               - - -
              - FERET: FacE REcognition -
              -               - - -
              - Labeled Face Parts in The Wild -
              -               - - -
              - Labeled Faces in The Wild -
              -               - - -
              - Unconstrained College Students -
              -               - - -
              - VGG Face 2 Dataset -
              -               - - -
              - YouTube Celebrities -
              -               - -
              -
              - - +
              +
              - - - - - - - - - - + \ No newline at end of file diff --git a/site/public/info/index.html b/site/public/info/index.html index 65510255..0b59e647 100644 --- a/site/public/info/index.html +++ b/site/public/info/index.html @@ -27,7 +27,7 @@
              -

              +

              Face Analysis

              Results are only stored for the duration of the analysis and are deleted when you leave this page.

              diff --git a/site/templates/home.html b/site/templates/home.html index 9756e21f..d5a2e59f 100644 --- a/site/templates/home.html +++ b/site/templates/home.html @@ -1,82 +1,39 @@ -{% extends 'layout.html' %} - -{% block content %} -
              -
              -
              -
              -
              -
              -
              - MegaPixels is an art project that explores the dark side of face recognition datasets and the future of computer vision. -
              - -
              - - -
              - -
              - Made by Adam Harvey in collaboration with Jules Laplace, and in partnership with Mozilla.
              - Read more about MegaPixels -
              -
              + + + + MegaPixels + + + + + + + + + + +
              + + +
              MegaPixels
              +               +
              + DATASETS + RESEARCH + ABOUT
              +
              +
              +
              - -
              -

              Face Recognition Datasets

              -
              -

              SUMMARY

              -
              -
              Found
              275 datasets
              -
              Created between
              1993-2018
              -
              Smallest dataset
              20 images
              -
              Largest dataset
              10,000,000 images
              -
              Highest resolution faces
              450x500 (Unconstrained College Students)
              -
              Lowest resolution faces
              16x20 pixels (QMUL SurvFace)
              -
              +
              +
              - -

              - MegaPixels is an online art project that explores the history of face recognition from the perspective of datasets. MegaPixels aims to unravel the meanings behind the data and expose the darker corners of the biometric industry that have contributed to its growth. -

              -

              - Through a mix of case studies, visualizations, and interactive tools, Megapixels will use face recognition datasets to tell the history of modern biometrics. Many people have contributed to the development of face recignition technology, both wittingly and unwittingly. Not only scientists, but also celebrities and regular internet users have played a part. -

              -

              - Face recognition is a mess of contradictinos. It works, yet it doesn't actually work. It's cheap and accessible, but also expensive and out of control. Face recognition research has achieved headline grabbing superhuman accuracies over 99.9%, yet in practice it's also dangerously inaccurate. -

              -

              - During a trial installation at Sudkreuz station in Berlin in 2018, 20% of the matches were wrong, a number so low that it should not have any connection to law enforcement or justice. And in London, the Metropolitan police had been using face recognition software that mistakenly identified an alarming 98% of people as criminals, which perhaps is a crime itself. -

              -
              - -
              -

              Dataset Portraits

              -

              - We have prepared detailed case studies of some of the more noteworthy datasets, including tools to help you learn what is contained in these datasets, and even whether your own face has been used to train these algorithms. -

              - -
              - {% for dataset in datasets %} - -
              - {{ dataset.title }} -
              -               - {% endfor %} +
              + MegaPixels ©2017-19 Adam R. Harvey /  + ahprojects.com
              -
              - -{% endblock %} - -{% block scripts %} - - - - - - - -{% endblock %} +
              + + + \ No newline at end of file -- cgit v1.2.3-70-g09d2 From 6c0046a05e3f96b7cade9afa238fa3c03f0e4f81 Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Thu, 28 Feb 2019 18:36:24 +0100 Subject: images: clear both --- site/assets/css/css.css | 1 + 1 file changed, 1 insertion(+) (limited to 'site') diff --git a/site/assets/css/css.css b/site/assets/css/css.css index 732386bd..2e73f625 100644 --- a/site/assets/css/css.css +++ b/site/assets/css/css.css @@ -355,6 +355,7 @@ section.images { flex-direction: row; align-items: flex-start; justify-content: center; + clear: both; } .image:only-child { width: 100%; -- cgit v1.2.3-70-g09d2 From e845766d970f4afefc2fc47367c3478413f98ff2 Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Thu, 28 Feb 2019 18:48:45 +0100 Subject: moz image --- site/assets/css/css.css | 4 ++-- site/content/pages/about/assets/mozilla.png | Bin 0 -> 10997 bytes site/content/pages/about/index.md | 2 +- site/public/about/index.html | 4 ++-- 4 files changed, 5 insertions(+), 5 deletions(-) create mode 100644 site/content/pages/about/assets/mozilla.png (limited to 'site') diff --git a/site/assets/css/css.css b/site/assets/css/css.css index 2e73f625..ab3bb4a7 100644 --- a/site/assets/css/css.css +++ b/site/assets/css/css.css @@ -404,8 +404,8 @@ section.fullwidth .image { } .sideimage img { margin-right: 10px; - width: 250px; - height: 250px; + width: 200px; + height: 200px; } /* blog index */ diff --git a/site/content/pages/about/assets/mozilla.png b/site/content/pages/about/assets/mozilla.png new file mode 100644 index 00000000..0fd4f115 Binary files /dev/null and b/site/content/pages/about/assets/mozilla.png differ diff --git a/site/content/pages/about/index.md b/site/content/pages/about/index.md index 66fac8ae..59f70d7a 100644 --- a/site/content/pages/about/index.md +++ b/site/content/pages/about/index.md @@ -36,4 +36,4 @@ MegaPixels aims to answer to these questions and reveal the stories behind the m ![sideimage:Jules LaPlace](assets/jules-laplace.jpg) **Jules LaPlace** is an American artist and technologist also based in Berlin. He was previously the CTO of a NYC digital agency and currently works at VFRAME, developing computer vision for human rights groups, and building creative software for artists. -**Mozilla** is a free software community founded in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions. The community is supported institutionally by the not-for-profit Mozilla Foundation and its tax-paying subsidiary, the Mozilla Corporation. +![sideimage:Mozilla](assets/mozilla.png) **Mozilla** is a free software community founded in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions. The community is supported institutionally by the not-for-profit Mozilla Foundation and its tax-paying subsidiary, the Mozilla Corporation. diff --git a/site/public/about/index.html b/site/public/about/index.html index 8583fd96..b7401ee8 100644 --- a/site/public/about/index.html +++ b/site/public/about/index.html @@ -39,8 +39,8 @@

              MegaPixels aims to answer to these questions and reveal the stories behind the millions of images used to train, evaluate, and power the facial recognition surveillance algorithms used today. MegaPixels is authored by Adam Harvey, developed in collaboration with Jules LaPlace, and produced in partnership with Mozilla.

            Adam Harvey

            Adam Harvey is an American artist and researcher based in Berlin. His previous projects (CV Dazzle, Stealth Wear, and SkyLift) explore the potential for countersurveillance as artwork. He is the founder of VFRAME (visual forensics software for human rights groups), the recipient of 2 PrototypeFund awards, and is currently a researcher in residence at Karlsruhe HfG studying artifical intelligence and datasets.

            Jules LaPlace

            Jules LaPlace is an American artist and technologist also based in Berlin. He was previously the CTO of a NYC digital agency and currently works at VFRAME, developing computer vision for human rights groups, and building creative software for artists.

            -

            Mozilla is a free software community founded in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions. The community is supported institutionally by the not-for-profit Mozilla Foundation and its tax-paying subsidiary, the Mozilla Corporation.

            -
            +
            Mozilla

            Mozilla is a free software community founded in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions. The community is supported institutionally by the not-for-profit Mozilla Foundation and its tax-paying subsidiary, the Mozilla Corporation.

            +