diff --git a/site/public/about/disclaimer/index.html b/site/public/about/disclaimer/index.html index 553bf084..5df5d656 100644 --- a/site/public/about/disclaimer/index.html +++ b/site/public/about/disclaimer/index.html @@ -18,10 +18,10 @@ The Darkside of Datasets

- Face Search - Datasets - Research - About + Face Search + Datasets + Research + About

diff --git a/site/public/about/index.html b/site/public/about/index.html index 363e8fc0..f1a28b0e 100644 --- a/site/public/about/index.html +++ b/site/public/about/index.html @@ -18,15 +18,16 @@ The Darkside of Datasets

- Face Search - Datasets - Research - About + Face Search + Datasets + Research + About

alt text

Credits

alt text

MegaPixels by Adam Harvey
Made with support from Mozilla
Site developed by Jules Laplace

The Darkside of Datasets

- Face Search - Datasets - Research - About + Face Search + Datasets + Research + About

diff --git a/site/public/about/privacy/index.html b/site/public/about/privacy/index.html index d1ec1c77..7ad9564f 100644 --- a/site/public/about/privacy/index.html +++ b/site/public/about/privacy/index.html @@ -18,10 +18,10 @@ The Darkside of Datasets

- Face Search - Datasets - Research - About + Face Search + Datasets + Research + About

@@ -84,7 +84,6 @@ megapixels.cc will take all steps reasonably necessary to ensure that your data

Disclosure Of Data

Legal Requirements

megapixels.cc may disclose your Personal Data in the good faith belief that such action is necessary to:

To comply with a legal obligation
To protect and defend the rights or property of megapixels.cc

The Darkside of Datasets

- Face Search - Datasets - Research - About + Face Search + Datasets + Research + About

Alt text here

Header 2

Style Examples

Alt text here

Header 1

Header 2

Header 3

Header 4

Header 5

diff --git a/site/public/about/terms/index.html b/site/public/about/terms/index.html index 4b9f4445..db8b9e57 100644 --- a/site/public/about/terms/index.html +++ b/site/public/about/terms/index.html @@ -18,10 +18,10 @@ The Darkside of Datasets

- Face Search - Datasets - Research - About + Face Search + Datasets + Research + About

diff --git a/site/public/datasets/lfw/index.html b/site/public/datasets/lfw/index.html index a130c24e..76549d25 100644 --- a/site/public/datasets/lfw/index.html +++ b/site/public/datasets/lfw/index.html @@ -18,28 +18,22 @@ The Darkside of Datasets

- Face Search - Datasets - Research - About + Face Search + Datasets + Research + About

Created 2007
Images 13,233
People 5,749
Created From Yahoo News images
Analyzed and searchable

Labeled Faces in The Wild is amongst the most widely used facial recognition training datasets in the world and is the first dataset of its kind to be created entirely from Internet photos. It includes 13,233 images of 5,749 people downloaded from the Internet, otherwise referred to as “The Wild”.

Labeled Faces in The Wild

Created

2007

Images

13,233

People

5,749

Created From

Yahoo News images

Search available

Searchable

Eight out of 5,749 people in the Labeled Faces in the Wild dataset. The face recognition training dataset is created entirely from photos downloaded from the Internet.

INTRO

It began in 2002. Researchers at University of Massachusetts Amherst were developing algorithms for facial recognition and they needed more data. Between 2002-2004 they scraped Yahoo News for images of public figures. Two years later they cleaned up the dataset and repackaged it as Labeled Faces in the Wild (LFW).

Since then the LFW dataset has become one of the most widely used datasets used for evaluating face recognition algorithms. The associated research paper “Labeled Faces in the Wild: A Database for Studying Face Recognition in Unconstrained Environments” has been cited 996 times reaching 45 different countries throughout the world.

The faces come from news stories and are mostly celebrities from the entertainment industry, politicians, and villains. It’s a sampling of current affairs and breaking news that has come to pass. The images, detached from their original context now server a new purpose: to train, evaluate, and improve facial recognition.

As the most widely used facial recognition dataset, it can be said that each individual in LFW has, in a small way, contributed to the current state of the art in facial recognition surveillance. John Cusack, Julianne Moore, Barry Bonds, Osama bin Laden, and even Moby are amongst these biometric pillars, exemplar faces provided the visual dimensions of a new computer vision future.

From Aaron Eckhart to Zydrunas Ilgauskas. A small sampling of the LFW dataset

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.

From Aaron Eckhart to Zydrunas Ilgauskas. A small sampling of the LFW dataset

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.

Usage

#!/usr/bin/python
 from matplotlib import plt
diff --git a/site/public/datasets/lfw/what/index.html b/site/public/datasets/lfw/what/index.html
new file mode 100644
index 00000000..52993a79
--- /dev/null
+++ b/site/public/datasets/lfw/what/index.html
@@ -0,0 +1,141 @@
+
+
+
+  MegaPixels
+  
+  
+  
+  
+  
+  
+  
+
+
+  
+    
+      
+      MegaPixels
+      The Darkside of Datasets
+    
+    
+      Face Search
+      Datasets
+      Research
+      About
+    
+  
+  
+    
+  Labeled Faces in The Wild
+
+Created 2007 (auto)
+Images 13,233 (auto)
+People 5,749 (auto)
+Created From Yahoo News images (auto)
+Analyzed and searchable (auto)
+
+Labeled Faces in The Wild is amongst the most widely used facial recognition training datasets in the world and is the first facial recognition dataset [^lfw_names_faces] of its kind to be created entirely from Internet photos. It includes 13,233 images of 5,749 people that appeared on Yahoo News between 2002 - 2004.
+
Eight out of 5,749 people in the Labeled Faces in the Wild dataset. The face recognition training dataset is created entirely from photos downloaded from the Internet.
INTRO
+It began in 2002. Researchers at University of Massachusetts Amherst were developing algorithms for facial recognition and they needed more data. Between 2002-2004 they scraped Yahoo News for images of public figures. Two years later they cleaned up the dataset and repackaged it as Labeled Faces in the Wild (LFW).
+Since then the LFW dataset has become one of the most widely used datasets used for evaluating face recognition algorithms. The associated research paper “Labeled Faces in the Wild: A Database for Studying Face Recognition in Unconstrained Environments” has been cited 996 times reaching 45 different countries throughout the world.
+The faces come from news stories and are mostly celebrities from the entertainment industry, politicians, and villains. It’s a sampling of current affairs and breaking news that has come to pass. The images, detached from their original context now server a new purpose: to train, evaluate, and improve facial recognition.
+As the most widely used facial recognition dataset, it can be said that each individual in LFW has, in a small way, contributed to the current state of the art in facial recognition surveillance. John Cusack, Julianne Moore, Barry Bonds, Osama bin Laden, and even Moby are amongst these biometric pillars, exemplar faces provided the visual dimensions of a new computer vision future.
+
From Aaron Eckhart to Zydrunas Ilgauskas. A small sampling of the LFW dataset
In addition to commercial use as an evaluation tool, all of the faces in LFW dataset are prepackaged into a popular machine learning code framework called scikit-learn.
+Usage
+#!/usr/bin/python
+from matplotlib import plt
+from sklearn.datasets import fetch_lfw_people
+lfw_people = fetch_lfw_people()
+lfw_person = lfw_people[0]
+plt.imshow(lfw_person)
+
+Commercial Use
+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
+load file: lfw_commercial_use.csv
+name_display,company_url,example_url,country,description
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Company Country Industries
Aratek China Biometric sensors for telecom, civil identification, finance, education, POS, and transportation
Aratek China Biometric sensors for telecom, civil identification, finance, education, POS, and transportation
Aratek China Biometric sensors for telecom, civil identification, finance, education, POS, and transportation
+Add 2-4 screenshots of companies mentioning LFW here
+
ReadSense
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:
+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. ¹.
+
+Citations
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Title Organization Country Type
3D-aided face recognition from videos University of Lyon France edu
A Community Detection Approach to Cleaning Extremely Large Face Database National University of Defense Technology, China China edu
+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.
+Notes
+According to BiometricUpdate.com², 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."
+
+
+"Chinese tourist town uses face recognition as an entry pass". New Scientist. November 17, 2016. https://www.newscientist.com/article/2113176-chinese-tourist-town-uses-face-recognition-as-an-entry-pass/↩
+"PING AN Tech facial recognition receives high score in latest LFW test results". https://www.biometricupdate.com/201702/ping-an-tech-facial-recognition-receives-high-score-in-latest-lfw-test-results↩
+
+
+
+
+  
+  
+    
+      MegaPixels.cc
+      Disclaimer
+      Terms of Use
+      Privacy
+      About
+      Team
+    
+    
+      MegaPixels ©2017-19 Adam R. Harvey / 
+      ahprojects.com
+    
+  
+
+
+
\ No newline at end of file
diff --git a/site/public/datasets/vgg_faces2/index.html b/site/public/datasets/vgg_faces2/index.html
index ee353047..95b5f7d7 100644
--- a/site/public/datasets/vgg_faces2/index.html
+++ b/site/public/datasets/vgg_faces2/index.html
@@ -18,23 +18,17 @@
       The Darkside of Datasets
     
     
-      Face Search
-      Datasets
-      Research
-      About
+      Face Search
+      Datasets
+      Research
+      About
     
   
   
     
-  
-Created 2007
-Images 13,233
-People 5,749
-Created From Yahoo News images
-Search available Searchable
-
-Labeled Faces in The Wild is amongst the most widely used facial recognition training datasets in the world and is the first dataset of its kind to be created entirely from Internet photos. It includes 13,233 images of 5,749 people downloaded from the Internet, otherwise referred to by researchers as “The Wild”.
-
Eight out of 5,749 people in the Labeled Faces in the Wild dataset. The face recognition training dataset is created entirely from photos downloaded from the Internet.
INTRO
+  Labeled Faces in The Wild
+
Created
2007
Images
13,233
People
5,749
Created From
Yahoo News images
Search available
[Searchable](#)
Labeled Faces in The Wild is amongst the most widely used facial recognition training datasets in the world and is the first dataset of its kind to be created entirely from Internet photos. It includes 13,233 images of 5,749 people downloaded from the Internet, otherwise referred to by researchers as “The Wild”.
+INTRO
 It began in 2002. Researchers at University of Massachusetts Amherst were developing algorithms for facial recognition and they needed more data. Between 2002-2004 they scraped Yahoo News for images of public figures. Two years later they cleaned up the dataset and repackaged it as Labeled Faces in the Wild (LFW).
 Since then the LFW dataset has become one of the most widely used datasets used for evaluating face recognition algorithms. The associated research paper “Labeled Faces in the Wild: A Database for Studying Face Recognition in Unconstrained Environments” has been cited 996 times reaching 45 different countries throughout the world.
 The faces come from news stories and are mostly celebrities from the entertainment industry, politicians, and villains. It’s a sampling of current affairs and breaking news that has come to pass. The images, detached from their original context now server a new purpose: to train, evaluate, and improve facial recognition.
diff --git a/site/public/index.html b/site/public/index.html
index ea3dc24c..3ce22936 100644
--- a/site/public/index.html
+++ b/site/public/index.html
@@ -18,23 +18,23 @@
       The Darkside of Datasets
     
     
-      Face Search
-      Datasets
-      Research
-      About
+      Face Search
+      Datasets
+      Research
+      About
     
   
   
     
   MegaPixels is an art project that explores the dark side of face recognition training data and the future of computer vision
 Made by Adam Harvey in partnership with Mozilla.

-Read more about MegaPixels
+Read more [about MegaPixels]
 [Explore Datasets]  [Explore Algorithms]
 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
 
-275 datasets found
+275 datsets found
 Created between the years 1993-2018
 Smallest dataset: 20 images
 Largest dataset: 10,000,000 images
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 90f142e9..55e02c6c 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
@@ -18,10 +18,10 @@
       The Darkside of Datasets
     
     
-      Face Search
-      Datasets
-      Research
-      About
+      Face Search
+      Datasets
+      Research
+      About
     
   
   
@@ -74,13 +74,6 @@
 
 
 
-  
-    MORE RESEARCH
-    
-      
-    
-  
-
   
   
     
diff --git a/site/public/research/index.html b/site/public/research/index.html
new file mode 100644
index 00000000..1f61dadf
--- /dev/null
+++ b/site/public/research/index.html
@@ -0,0 +1,50 @@
+
+
+
+  MegaPixels
+  
+  
+  
+  
+  
+  
+  
+
+
+  
+    
+      
+      MegaPixels
+      The Darkside of Datasets
+    
+    
+      Face Search
+      Datasets
+      Research
+      About
+    
+  
+  
+    
+  Research Blog
+The darkside of datasets and the future of computer vision
+
From 1 to 100 Pixels
Photographs are for romantics. For the rest of us, it's all about data.  And a photo contains a massive amount of information about who you are.
+
+  
+  
+    
+      MegaPixels.cc
+      Disclaimer
+      Terms of Use
+      Privacy
+      About
+      Team
+    
+    
+      MegaPixels ©2017-19 Adam R. Harvey / 
+      ahprojects.com
+    
+  
+
+
+
\ No newline at end of file
diff --git a/site/templates/layout.html b/site/templates/layout.html
index 5b5833be..7558163e 100644
--- a/site/templates/layout.html
+++ b/site/templates/layout.html
@@ -18,10 +18,10 @@
       The Darkside of Datasets
     
     
-      Face Search
-      Datasets
-      Research
-      About
+      Face Search
+      Datasets
+      Research
+      About
     
   
   
diff --git a/site/templates/research.html b/site/templates/research.html
index 22e494c2..0bb9fa30 100644
--- a/site/templates/research.html
+++ b/site/templates/research.html
@@ -20,16 +20,4 @@
   
 
   {{ content }}
-
-  
-    MORE RESEARCH
-    
-      {% for blogpost in blogposts %}
-        
-          {{ blogpost.title }}
-          {{ blogpost.date }}
-        
-      {% endfor %}
-    
-  
 {% endblock %}
-- 
cgit v1.2.3-70-g09d2


From 2d950c3fa3b8107f941a80f88127ab45e371d128 Mon Sep 17 00:00:00 2001
From: Jules Laplace 
Date: Thu, 6 Dec 2018 19:39:29 +0100
Subject: homepage css

---
 builder/README.md                                  |   3 +
 builder/builder.py                                 |  10 +-
 builder/parser.py                                  |   4 +-
 site/assets/css/css.css                            |  93 ++++++++-
 site/assets/js/app/face.js                         | 213 +++++++++++++++++++++
 site/assets/js/app/site.js                         |   3 +-
 site/public/about/credits/index.html               |   1 +
 site/public/about/disclaimer/index.html            |   1 +
 site/public/about/index.html                       |   1 +
 site/public/about/press/index.html                 |   1 +
 site/public/about/privacy/index.html               |   1 +
 site/public/about/style/index.html                 |   3 +-
 site/public/about/terms/index.html                 |   1 +
 site/public/datasets/lfw/index.html                | 160 +++++++++++++---
 site/public/datasets/lfw/what/index.html           |   1 +
 site/public/datasets/vgg_faces2/index.html         |   1 +
 site/public/index.html                             |  42 ++--
 .../research/01_from_1_to_100_pixels/index.html    |   1 +
 site/public/research/index.html                    |   1 +
 site/templates/home.html                           |  32 ++++
 site/templates/layout.html                         |   1 +
 21 files changed, 524 insertions(+), 50 deletions(-)
 create mode 100644 site/assets/js/app/face.js
 create mode 100644 site/templates/home.html

(limited to 'site/public/index.html')

diff --git a/builder/README.md b/builder/README.md
index 1a6d3a1e..57c024cb 100644
--- a/builder/README.md
+++ b/builder/README.md
@@ -19,3 +19,6 @@ authors: Adam Harvey, Berit Gilma, Matthew Stender
 
 Static assets: `v1/site/about/assets/picture.jpg`
 Dataset assets: `v1/datasets/lfw/assets/picture.jpg`
+
+## Markup
+
diff --git a/builder/builder.py b/builder/builder.py
index 0e404b88..620fc710 100644
--- a/builder/builder.py
+++ b/builder/builder.py
@@ -29,10 +29,12 @@ def build_page(fn, research_posts):
   output_path = public_path + metadata['url']
   output_fn = os.path.join(output_path, "index.html")
 
-  is_research = False
+  skip_h1 = False
 
-  if 'research/' in fn:
-    is_research = True
+  if metadata['url'] == '/':
+    template = env.get_template("home.html")
+  elif 'research/' in fn:
+    skip_h1 = True
     template = env.get_template("research.html")
   else:
     template = env.get_template("page.html")
@@ -47,7 +49,7 @@ def build_page(fn, research_posts):
   if 'index.md' in fn:
     s3.sync_directory(dirname, s3_dir, metadata)
 
-  content = parser.parse_markdown(sections, s3_path, skip_h1=is_research)
+  content = parser.parse_markdown(sections, s3_path, skip_h1=skip_h1)
 
   html = template.render(
     metadata=metadata,
diff --git a/builder/parser.py b/builder/parser.py
index da3044a0..dd3643bf 100644
--- a/builder/parser.py
+++ b/builder/parser.py
@@ -46,15 +46,13 @@ def format_metadata(section):
 def parse_markdown(sections, s3_path, skip_h1=False):
   groups = []
   current_group = []
-  seen_metadata = False
   for section in sections:
     if skip_h1 and section.startswith('# '):
       continue
-    elif section.startswith('+ ') and not seen_metadata:
+    elif section.startswith('+ '):
       groups.append(format_section(current_group, s3_path))
       groups.append(format_metadata(section))
       current_group = []
-      seen_metadata = True
     elif '![wide:' in section:
       groups.append(format_section(current_group, s3_path))
       groups.append(format_section([section], s3_path, type='wide'))
diff --git a/site/assets/css/css.css b/site/assets/css/css.css
index 843809a8..9ac35699 100644
--- a/site/assets/css/css.css
+++ b/site/assets/css/css.css
@@ -5,9 +5,11 @@ html, body {
   width: 100%;
   min-height: 100%;
   font-family: 'Roboto', sans-serif;
-  background: #191919;
   color: #b8b8b8;
 }
+html {
+  background: #191919;
+}
 
 /* header */
 
@@ -119,12 +121,14 @@ h1 {
   font-size: 24pt;
   margin: 75px 0 10px;
   padding: 0;
+  transition: color 0.2s cubic-bezier(0,0,1,1);
 }
 h2, h3 {
   margin: 0 0 20px 0;
   padding: 0;
   font-size: 11pt;
   font-weight: 500;
+  transition: color 0.2s cubic-bezier(0,0,1,1);
 }
 
 th, .gray, h2, h3 {
@@ -281,6 +285,9 @@ section.wide .image {
   max-width: 620px;
   margin: 10px auto 0 auto;
 }
+
+/* blog index */
+
 .research_index {
   margin-top: 40px;
 }
@@ -289,10 +296,88 @@ section.wide .image {
 }
 .research_index h1 {
   margin-top: 20px;
+  text-decoration: underline;
+}
+.desktop .research_index section:hover h1 {
+  color: #fff;
+}
+.research_index section:hover h2 {
+  color: #ddd;
 }
 
-/* blogpost index */
+/* home page */
 
-.blogposts div {
-  margin-bottom: 5px;
+.hero {
+  position: relative;
+  width: 100%;
+  max-width: 1200px;
+  height: 50vw;
+  max-height: 70vh;
+  display: flex;
+  align-items: center;
+  margin: 0 auto;
+}
+#face_container {
+  pointer-events: none;
+  position: absolute;
+  width: 50vw;
+  height: 50vw;
+  max-height: 70vh;
+  top: 0;
+  right: 0;
+  z-index: -1;
+  text-align: center;
+}
+.currentFace {
+  position: absolute;
+  bottom: 50px;
+  width: 100%;
+  left: 0;
+  text-align: center;
+}
+.intro {
+  max-width: 640px;
+  padding: 75px 0 75px 10px;
+  z-index: 1;
+}
+.intro .headline {
+  font-family: 'Roboto Mono', monospace;
+  font-size: 16pt;  
+}
+.intro .buttons {
+  margin: 40px 0;
+}
+.intro button {
+  font-family: 'Roboto', sans-serif;
+  padding: 8px 12px;
+  border-radius: 6px;
+  border: 1px solid transparent;
+  cursor: pointer;
+  font-size: 11pt;
+  margin-right: 10px;
+  transition: color 0.1s cubic-bezier(0,0,1,1), background-color 0.1s cubic-bezier(0,0,1,1);
+}
+.intro button.normal {
+  background: #191919;
+  border-color: #444;
+  color: #ddd;
+}
+.intro button.important {
+  background: #444;
+  border-color: #444;
+  color: #ddd;
+}
+.desktop .intro button:hover {
+  background: #666;
+  border-color: #666;
+  color: #fff;
+}
+.intro .under {
+  color: #888;
+}
+.intro .under a {
+  color: #bbb;
+}
+.desktop .intro .under a:hover {
+  color: #fff;
 }
\ No newline at end of file
diff --git a/site/assets/js/app/face.js b/site/assets/js/app/face.js
new file mode 100644
index 00000000..e8bcd313
--- /dev/null
+++ b/site/assets/js/app/face.js
@@ -0,0 +1,213 @@
+var face = (function(){
+  var container = document.querySelector("#face_container")
+  var camera, controls, scene, renderer
+  var mouse = new THREE.Vector2(0.5, 0.5)
+  var mouseTarget = new THREE.Vector2(0.5, 0.5)
+  var POINT_SCALE = 1.8
+  var FACE_POINT_COUNT = 68
+  var SWAP_TIME = 500
+  var cubes = [], meshes = []
+  var currentFace = document.querySelector('.currentFace')
+  var faceBuffer = (function () {
+    var a = new Array(FACE_POINT_COUNT)
+    for (let i = 0; i < FACE_POINT_COUNT; i++) {
+      a[i] = new THREE.Vector3()
+    }
+    return a
+  })()
+  var last_t = 0, start_t = 0
+  var colors = [
+    0xff3333,
+    0xff8833,
+    0xffff33,
+    0x338833,
+    0x3388ff,
+    0x3333ff,
+    0x8833ff,
+    0xff3388,
+    0xffffff,
+  ]
+  var swapping = false, swap_count = 0, swapFrom, swapTo, face_names, faces
+  init()
+
+  function init() {
+    fetch("/assets/data/3dlm_0_10.json")
+    .then(req => req.json())
+    .then(data => {
+      face_names = Object.keys(data)
+      faces = face_names.map(name => recenter(data[name]))
+      setup()
+      build(faces[0])
+      updateFace(faces[0])
+      setCurrentFace(face_names[0])
+      swapTo = faces[0]
+      animate()
+    })
+  }
+  function setup() {
+    var w = window.innerWidth / 2
+    var h = Math.min(window.innerWidth / 2, window.innerHeight * 0.7)
+    camera = new THREE.PerspectiveCamera(70, w/h, 1, 10000)
+    camera.position.x = 0
+    camera.position.y = 0
+    camera.position.z = 250
+
+    scene = new THREE.Scene()
+    scene.background = new THREE.Color(0x191919)
+
+    renderer = new THREE.WebGLRenderer({ antialias: true })
+    renderer.setPixelRatio(window.devicePixelRatio)
+    renderer.setSize(w, h)
+    container.appendChild(renderer.domElement)
+    document.body.addEventListener('mousemove', onMouseMove)
+    // renderer.domElement.addEventListener('mousedown', swap)
+    setInterval(swap, 5000)
+  }
+  function build(points) {
+    var matrix = new THREE.Matrix4()
+    var quaternion = new THREE.Quaternion()
+
+    for (var i = 0; i < FACE_POINT_COUNT; i++) {
+      var p = points[i]
+      var geometry = new THREE.BoxBufferGeometry()
+      var position = new THREE.Vector3(p[0], p[1], p[2])
+      var rotation = new THREE.Euler()
+      var scale = new THREE.Vector3()
+      var color = new THREE.Color()
+      scale.x = scale.y = scale.z = POINT_SCALE
+      quaternion.setFromEuler(rotation, false)
+      matrix.compose(position, quaternion, scale)
+      geometry.applyMatrix(matrix)
+      material = new THREE.MeshBasicMaterial({ color: color.setHex(0xffffff) })
+      cube = new THREE.Mesh(geometry, material)
+      scene.add(cube)
+      cubes.push(cube)
+    }
+
+    meshes = getLineGeometry(points).map((geometry, i) => {
+      var color = new THREE.Color()
+      var material = new MeshLineMaterial({
+        color: color.setHex(colors[i % colors.length]),
+      })
+      var line = new MeshLine()
+      line.setGeometry(geometry, _ => 1.5)
+      var mesh = new THREE.Mesh(line.geometry, material)
+      mesh.geometry.dynamic = true
+      scene.add(mesh)
+      return [line, mesh]
+    })
+  }
+  function lerpPoints(n, A, B, C) {
+    for (let i = 0, len = A.length; i < len; i++) {
+      lerpPoint(n, A[i], B[i], C[i])
+    }
+  }
+  function lerpPoint(n, A, B, C) {
+    C.x = lerp(n, A.x, B.x)
+    C.y = lerp(n, A.y, B.y)
+    C.z = lerp(n, A.z, B.z)
+  }
+  function lerp(n, a, b) {
+    return (b-a) * n + a
+  }
+  function swap(){
+    if (swapping) return
+    start_t = last_t
+    swapping = true
+    swap_count = (swap_count + 1) % faces.length
+    swapFrom = swapTo
+    swapTo = faces[swap_count]
+    setCurrentFace(face_names[swap_count])
+  }
+  function setCurrentFace(name) {
+    name = name.replace('.png', '').split('_').filter(s => !s.match(/\d+/)).join(' ')
+    currentFace.innerHTML = name
+  }
+  function update_swap(t){
+    var n = (t - start_t) / SWAP_TIME
+    if (n > 1) {
+      swapping = false
+      n = 1
+    }
+    lerpPoints(n, swapFrom, swapTo, faceBuffer)
+    updateFace(faceBuffer)
+  }
+  function updateFace(points) {
+    updateCubeGeometry(points)
+    updateLineGeometry(points)
+  }
+  function updateCubeGeometry(points) {
+    cubes.forEach((cube, i) => {
+      const p = points[i]
+      cube.position.set(p.x, p.y, p.z)
+    })
+  }
+  function updateLineGeometry(points) {
+    getLineGeometry(points).map((geometry, i) => {
+      var [line, mesh] = meshes[i]
+      line.setGeometry(geometry, _ => 1.5)
+      mesh.geometry.vertices = line.geometry.vertices
+      mesh.geometry.verticesNeedUpdate = true
+    })
+  }
+  function getLineGeometry(points) {
+    return [
+      points.slice(0, 17),
+      points.slice(17, 22),
+      points.slice(22, 27),
+      points.slice(27, 31),
+      points.slice(31, 36),
+      points.slice(36, 42),
+      points.slice(42, 48),
+      points.slice(48)
+    ].map((a, i) => {
+      var geometry = new THREE.Geometry()
+      a.forEach(p => geometry.vertices.push(p))
+      if (i > 4) {
+        geometry.vertices.push(a[0])
+      }
+      return geometry
+    })
+  }
+  function getBounds(obj) {
+    return obj.reduce((a, p) => {
+      return [
+        Math.min(a[0], p[0]),
+        Math.max(a[1], p[0]),
+        Math.min(a[2], p[1]),
+        Math.max(a[3], p[1]),
+        Math.min(a[4], p[2]),
+        Math.max(a[5], p[2]),
+      ]
+    }, [Infinity, -Infinity, Infinity, -Infinity, Infinity, -Infinity])
+  }
+  function recenter(obj) {
+    const bounds = getBounds(obj)
+    const x_width = (bounds[1] - bounds[0]) / 2
+    const y_width = (bounds[3] - bounds[2]) / -3
+    const z_width = (bounds[5] - bounds[4]) / 2
+    return obj.map(p => {
+      p[0] = p[0] - bounds[0] - x_width
+      p[1] = -p[1] + bounds[1] + y_width
+      p[2] = p[2] - bounds[2] + z_width
+      return new THREE.Vector3(p[0], p[1], p[2])
+    })
+  }
+  //
+  function onMouseMove(e) {
+    mouse.x = e.clientX / window.innerWidth
+    mouse.y = e.clientY / window.innerHeight
+  }
+  function animate(t) {
+    requestAnimationFrame(animate)
+    if (swapping) update_swap(t)
+    renderer.render(scene, camera)
+    scene.rotation.y += 0.01 * Math.PI
+    mouseTarget.x += (mouse.x - mouseTarget.x) * 0.1
+    mouseTarget.y += (mouse.y - mouseTarget.y) * 0.1
+    scene.rotation.x = (mouseTarget.y - 0.5) * Math.PI / 2
+    // scene.rotation.y = (mouseTarget.x - 0.5) * Math.PI
+    scene.rotation.y += 0.01
+    last_t = t
+  }
+})()
diff --git a/site/assets/js/app/site.js b/site/assets/js/app/site.js
index 12bee3ec..eb6886c2 100644
--- a/site/assets/js/app/site.js
+++ b/site/assets/js/app/site.js
@@ -7,7 +7,8 @@ const isDesktop = !isMobile
 const htmlClassList = document.body.parentNode.classList
 htmlClassList.add(isDesktop ? 'desktop' : 'mobile')
 
-function toArray(A) { return Array.prototype.slice.apply(A) }
+function toArray(a) { return Array.prototype.slice.apply(a) }
+function choice(a) { return a[Math.floor(Math.random()*a.length)]}
 
 var site = (function(){
   var site = {}
diff --git a/site/public/about/credits/index.html b/site/public/about/credits/index.html
index f1a28b0e..65bc7ac4 100644
--- a/site/public/about/credits/index.html
+++ b/site/public/about/credits/index.html
@@ -52,5 +52,6 @@
     
   
 
+
 
 
\ No newline at end of file
diff --git a/site/public/about/disclaimer/index.html b/site/public/about/disclaimer/index.html
index 5df5d656..b0215bde 100644
--- a/site/public/about/disclaimer/index.html
+++ b/site/public/about/disclaimer/index.html
@@ -52,5 +52,6 @@

Company	Country	Industries
Aratek	China	Biometric sensors for telecom, civil identification, finance, education, POS, and transportation
Aratek	China	Biometric sensors for telecom, civil identification, finance, education, POS, and transportation
Aratek	China	Biometric sensors for telecom, civil identification, finance, education, POS, and transportation

Title	Organization	Country	Type
3D-aided face recognition from videos	University of Lyon	France	edu
A Community Detection Approach to Cleaning Extremely Large Face Database	National University of Defense Technology, China	China	edu


   
 
+
 
 
\ No newline at end of file
diff --git a/site/public/about/index.html b/site/public/about/index.html
index f1a28b0e..65bc7ac4 100644
--- a/site/public/about/index.html
+++ b/site/public/about/index.html
@@ -52,5 +52,6 @@


   
 
+
 
 
\ No newline at end of file
diff --git a/site/public/about/press/index.html b/site/public/about/press/index.html
index e5763036..09c89165 100644
--- a/site/public/about/press/index.html
+++ b/site/public/about/press/index.html
@@ -50,5 +50,6 @@


   
 
+
 
 
\ No newline at end of file
diff --git a/site/public/about/privacy/index.html b/site/public/about/privacy/index.html
index 7ad9564f..5675f072 100644
--- a/site/public/about/privacy/index.html
+++ b/site/public/about/privacy/index.html
@@ -129,5 +129,6 @@ You are advised to review this Privacy Policy periodically for any changes. Chan


   
 
+
 
 
\ No newline at end of file
diff --git a/site/public/about/style/index.html b/site/public/about/style/index.html
index eea861ac..f2c0d4b8 100644
--- a/site/public/about/style/index.html
+++ b/site/public/about/style/index.html
@@ -27,7 +27,7 @@
   
     
   Style Examples
-
Alt text here
Header 1
+
Alt text here
Date
17-Jan-2019
Numbers
17
Identities
12,139
But also
This is a test of the stylesheet
Header 1
 Header 2
 Header 3
 Header 4
@@ -85,5 +85,6 @@ But let's throw in a <b>tag</b>.
     
   
 
+
 
 
\ No newline at end of file
diff --git a/site/public/about/terms/index.html b/site/public/about/terms/index.html
index db8b9e57..078c339f 100644
--- a/site/public/about/terms/index.html
+++ b/site/public/about/terms/index.html
@@ -64,5 +64,6 @@


   
 
+
 
 
\ No newline at end of file
diff --git a/site/public/datasets/lfw/index.html b/site/public/datasets/lfw/index.html
index 76549d25..39052b44 100644
--- a/site/public/datasets/lfw/index.html
+++ b/site/public/datasets/lfw/index.html
@@ -27,23 +27,22 @@
   
     
   Labeled Faces in The Wild
-
Created
2007
Images
13,233
People
5,749
Created From
Yahoo News images
Search available
Searchable
Labeled Faces in The Wild is amongst the most widely used facial recognition training datasets in the world and is the first dataset of its kind to be created entirely from Internet photos. It includes 13,233 images of 5,749 people downloaded from the Internet, otherwise referred to as “The Wild”.
-
Eight out of 5,749 people in the Labeled Faces in the Wild dataset. The face recognition training dataset is created entirely from photos downloaded from the Internet.
INTRO
+
Created
2007
Images
13,233
People
5,749
Created From
Yahoo News images
Search available
Searchable
Labeled Faces in The Wild (LFW) is amongst the most widely used facial recognition training datasets in the world and is the first of its kind to be created entirely from images that were 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.
+{INSERT IMAGE SEARCH MODULE}
+{INSERT TEXT SEARCH MODULE}
+
Eight out of 5,749 people in the Labeled Faces in the Wild dataset. The face recognition training dataset is created entirely from photos downloaded from the Internet.
INTRO
 It began in 2002. Researchers at University of Massachusetts Amherst were developing algorithms for facial recognition and they needed more data. Between 2002-2004 they scraped Yahoo News for images of public figures. Two years later they cleaned up the dataset and repackaged it as Labeled Faces in the Wild (LFW).
 Since then the LFW dataset has become one of the most widely used datasets used for evaluating face recognition algorithms. The associated research paper “Labeled Faces in the Wild: A Database for Studying Face Recognition in Unconstrained Environments” has been cited 996 times reaching 45 different countries throughout the world.
 The faces come from news stories and are mostly celebrities from the entertainment industry, politicians, and villains. It’s a sampling of current affairs and breaking news that has come to pass. The images, detached from their original context now server a new purpose: to train, evaluate, and improve facial recognition.
 As the most widely used facial recognition dataset, it can be said that each individual in LFW has, in a small way, contributed to the current state of the art in facial recognition surveillance. John Cusack, Julianne Moore, Barry Bonds, Osama bin Laden, and even Moby are amongst these biometric pillars, exemplar faces provided the visual dimensions of a new computer vision future.
-
From Aaron Eckhart to Zydrunas Ilgauskas. A small sampling of the LFW dataset
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.
-Usage
-#!/usr/bin/python
-from matplotlib import plt
-from sklearn.datasets import fetch_lfw_people
-lfw_people = fetch_lfw_people()
-lfw_person = lfw_people[0]
-plt.imshow(lfw_person)
-
+
The entire LFW dataset cropped to facial regions
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.
+Facts
+The person with the most images is:
+The person with the least images is:
 Commercial Use
-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
+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]."
 load file: lfw_commercial_use.csv
 name_display,company_url,example_url,country,description
 
@@ -73,11 +72,24 @@ name_display,company_url,example_url,country,description
 
 
 Add 2-4 screenshots of companies mentioning LFW here
-
ReadSense
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.
+
 "PING AN Tech facial recognition receives high score in latest LFW test results"
+ "Face Recognition Performance in LFW benchmark"
+ "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:
-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. ¹.
-
 Citations
+Overall, LFW has at least 456 citations from 123 countries. Sed ut perspiciatis, unde omnis iste natus error sit voluptatem accusantium doloremque laudantium, totam rem aperiam eaque ipsa, quae ab illo inventore veritatis et quasi architecto beatae vitae dicta sunt, explicabo. Nemo enim ipsam voluptatem, quia voluptas sit, aspernatur aut odit aut fugit, sed quia consequuntur magni dolores eos.
+Sed ut perspiciatis, unde omnis iste natus error sit voluptatem accusantium doloremque laudantium, totam rem aperiam eaque ipsa, quae ab illo inventore veritatis et quasi architecto beatae vitae dicta sunt, explicabo. Nemo enim ipsam voluptatem, quia voluptas sit, aspernatur aut odit aut fugit, sed quia consequuntur magni dolores eos.
+
Distribution of citations per year per country for the top 5 countries with citations for the LFW Dataset
Geographic distributions of citations for the LFW Dataset
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.
+Right to Removal
+If you are affected by disclosure of your identity in this dataset please do contact the authors, many state that they are willing to remove images upon request. The authors of the LFW can be reached from the emails posted in their paper:
+You can use the following message to request removal from the dataset:
+Dear [researcher name],
+I am writing to you about the "LFW Dataset". Recently I have discovered that your dataset includes my identity and no longer wish to be included in your dataset
+MegaPixels is an educational art project developed for academic purposes. In no way does this project aim to villify the researchers who produced the datasets. The aim of this project is to encourage discourse around ethics and consent in artificial intelligence by providing information about these datasets that is otherwise difficult to obtain or inaccessible to other researchers.
+Supplementary Data
+Sed ut perspiciatis, unde omnis iste natus error sit voluptatem accusantium doloremque laudantium, totam rem aperiam eaque ipsa, quae ab illo inventore veritatis et quasi architecto beatae vitae dicta sunt, explicabo. Nemo enim ipsam voluptatem, quia voluptas sit, aspernatur aut odit aut fugit, sed quia consequuntur magni dolores eos.
 
@@ -99,18 +111,119 @@ name_display,company_url,example_url,country,description
 
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
 
 Title China
 edu
 
3D-aided face recognition from videos University of Lyon France edu
3D-aided face recognition from videos University of Lyon France edu
3D-aided face recognition from videos University of Lyon France edu
3D-aided face recognition from videos University of Lyon France edu
3D-aided face recognition from videos University of Lyon France edu
3D-aided face recognition from videos University of Lyon France edu
3D-aided face recognition from videos University of Lyon France edu
3D-aided face recognition from videos University of Lyon France edu
3D-aided face recognition from videos University of Lyon France edu
3D-aided face recognition from videos University of Lyon France edu
3D-aided face recognition from videos University of Lyon France edu
3D-aided face recognition from videos University of Lyon France edu
3D-aided face recognition from videos University of Lyon France edu
 
 
-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.
-Notes
-According to BiometricUpdate.com², 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."
+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('{:,} images at {}x{}'.format(n_samples, w, h))
+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)
+
 
 
-"Chinese tourist town uses face recognition as an entry pass". New Scientist. November 17, 2016. https://www.newscientist.com/article/2113176-chinese-tourist-town-uses-face-recognition-as-an-entry-pass/↩
-"PING AN Tech facial recognition receives high score in latest LFW test results". https://www.biometricupdate.com/201702/ping-an-tech-facial-recognition-receives-high-score-in-latest-lfw-test-results↩
-
+
 
 
 
@@ -130,5 +243,6 @@ name_display,company_url,example_url,country,description
     
   
 
+
 
 
\ No newline at end of file
diff --git a/site/public/datasets/lfw/what/index.html b/site/public/datasets/lfw/what/index.html
index 52993a79..ceafb35a 100644
--- a/site/public/datasets/lfw/what/index.html
+++ b/site/public/datasets/lfw/what/index.html
@@ -137,5 +137,6 @@ name_display,company_url,example_url,country,description


   
 
+
 
 
\ No newline at end of file
diff --git a/site/public/datasets/vgg_faces2/index.html b/site/public/datasets/vgg_faces2/index.html
index 95b5f7d7..3f778f71 100644
--- a/site/public/datasets/vgg_faces2/index.html
+++ b/site/public/datasets/vgg_faces2/index.html
@@ -58,5 +58,6 @@


   
 
+
 
 
\ No newline at end of file
diff --git a/site/public/index.html b/site/public/index.html
index 3ce22936..51006b59 100644
--- a/site/public/index.html
+++ b/site/public/index.html
@@ -26,22 +26,31 @@
   
   
     
-  MegaPixels is an art project that explores the dark side of face recognition training data and the future of computer vision
-Made by Adam Harvey in partnership with Mozilla.

-Read more [about MegaPixels]
-[Explore Datasets]  [Explore Algorithms]
-Facial Recognition Datasets
+  
+    
+      
+    
+    
+      
+        MegaPixels is an art project that explores the dark side of face recognition and the future of computer vision.
+      
+
+      
+        
+      
+
+      
+        Made by Adam Harvey in partnership with Mozilla.

+        Read more about MegaPixels
+      
+    
+  
+
+  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
-
-275 datsets found
-Created between the years 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)

+
 
   
   
@@ -59,5 +68,10 @@ Read more [about MegaPixels]

Labeled Faces in The Wild

Created

2007

Images

13,233

People

5,749

Created From

Yahoo News images

Search available

Searchable

Labeled Faces in The Wild (LFW) is amongst the most widely used facial recognition training datasets in the world and is the first of its kind to be created entirely from images that were 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.

Labeled Faces in the Wild

Created

2007

Images

13,233

People

5,749

Created From

Yahoo News images

Search available

Searchable

Labeled Faces in The Wild (LFW) is amongst 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.

{INSERT IMAGE SEARCH MODULE}

{INSERT TEXT SEARCH MODULE}

Eight out of 5,749 people in the Labeled Faces in the Wild dataset. The face recognition training dataset is created entirely from photos downloaded from the Internet.

INTRO

The entire LFW dataset cropped to facial regions

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.

Facts

The person with the most images is: -The person with the least images is:

Commercial Use

load file: lfw_names_gender_kg_min.csv
+Name, Images, Gender, Description
+

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

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.

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]

need citations

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 (need citation)

former President George W. Bush

Colin Powel (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.

load file: lfw_commercial_use.csv
 name_display,company_url,example_url,country,description

@@ -83,13 +110,18 @@ name_display,company_url,example_url,country,description

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.

Right to Removal

If you are affected by disclosure of your identity in this dataset please do contact the authors, many state that they are willing to remove images upon request. The authors of the LFW can be reached from the emails posted in their paper:

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 "LFW Dataset". Recently I have discovered that your dataset includes my identity and no longer wish to be included in your dataset

MegaPixels is an educational art project developed for academic purposes. In no way does this project aim to villify the researchers who produced the datasets. The aim of this project is to encourage discourse around ethics and consent in artificial intelligence by providing information about these datasets that is otherwise difficult to obtain or inaccessible to other researchers.

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]

Supplementary Data

Sed ut perspiciatis, unde omnis iste natus error sit voluptatem accusantium doloremque laudantium, totam rem aperiam eaque ipsa, quae ab illo inventore veritatis et quasi architecto beatae vitae dicta sunt, explicabo. Nemo enim ipsam voluptatem, quia voluptas sit, aspernatur aut odit aut fugit, sed quia consequuntur magni dolores eos.

Researchers, journ

@@ -221,6 +253,9 @@ imageio.imwrite('lfw_montage_full.png', montage) montage_960 = imutils.resize(montage, width=960) imageio.imwrite('lfw_montage_960.jpg', montage_960) +

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

diff --git a/site/public/datasets/vgg_face2/index.html b/site/public/datasets/vgg_face2/index.html new file mode 100644 index 00000000..24a1059b --- /dev/null +++ b/site/public/datasets/vgg_face2/index.html @@ -0,0 +1,84 @@ + + + + MegaPixels + + + + + + + + + +

+ +

MegaPixels

+ The Darkside of Datasets + +

+ Face Search + 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.

{INSERT IMAGE SEARCH MODULE}

{INSERT TEXT SEARCH MODULE}

load file: lfw_names_gender_kg_min.csv
+Name, Images, Gender, Description
+

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 51006b59..91ff467a 100644 --- a/site/public/index.html +++ b/site/public/index.html @@ -71,6 +71,7 @@ + diff --git a/site/public/research/00_introduction/index.html b/site/public/research/00_introduction/index.html new file mode 100644 index 00000000..8f598f5b --- /dev/null +++ b/site/public/research/00_introduction/index.html @@ -0,0 +1,86 @@ + + + + MegaPixels + + + + + + + + + +

+ +

MegaPixels

+ The Darkside of Datasets + +

+ Face Search + Datasets + Research + About +

+ +

Untitled Page

Posted

2018-12-31

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.

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]."

One problem with FERET dataset was that the photos were in controlled settings. For face recognition to work it would have to be used in uncontrolled settings. Even newer datasets such as the Multi-PIE (Pose, Illumination, and Expression) from Carnegie Mellon University included only indoor photos of cooperative subjects. Not only were the photos completely unrealistic, CMU's Multi-Pie included only 18 individuals and cost $500 for academic use [^cmu_multipie_cost], took years to create, and required consent from every participant.

Sharman, Jon. "Metropolitan Police's facial recognition technology 98% inaccurate, figures show". 2018. https://www.independent.co.uk/news/uk/home-news/met-police-facial-recognition-success-south-wales-trial-home-office-false-positive-a8345036.html ↩
Calle, Dan. "Supercomptuers". 1997. http://ei.cs.vt.edu/~history/SUPERCOM.Calle.HTML ↩

+ +

+ + + + + \ No newline at end of file diff --git a/site/public/research/index.html b/site/public/research/index.html index cf9546e1..59a5fee9 100644 --- a/site/public/research/index.html +++ b/site/public/research/index.html @@ -28,7 +28,7 @@

Research Blog

The darkside of datasets and the future of computer vision

From 1 to 100 Pixels

Photographs are for romantics. For the rest of us, it's all about data. And a photo contains a massive amount of information about who you are.

Untitled Page

From 1 to 100 Pixels

Photographs are for romantics. For the rest of us, it's all about data. And a photo contains a massive amount of information about who you are.

Title