diff options
Diffstat (limited to 'site/public/datasets/ijb_c/index.html')
| -rw-r--r-- | site/public/datasets/ijb_c/index.html | 57 |
1 files changed, 56 insertions, 1 deletions
diff --git a/site/public/datasets/ijb_c/index.html b/site/public/datasets/ijb_c/index.html index 83d572ab..06232f72 100644 --- a/site/public/datasets/ijb_c/index.html +++ b/site/public/datasets/ijb_c/index.html @@ -74,7 +74,62 @@ <div class='gray'>Website</div> <div><a href='https://www.nist.gov/programs-projects/face-challenges' target='_blank' rel='nofollow noopener'>nist.gov</a></div> </div></div><p>[ page under development ]</p> -<p>The IARPA Janus Benchmark C is a dataset created by</p> +<p>The IARPA Janus Benchmark C (IJB–C) is a dataset of web images used for face recognition research and development. The IJB–C dataset contains 3,531 people</p> +<p>Among the target list of 3,531 names are activists, artists, journalists, foreign politicians,</p> +<ul> +<li>Subjects 3531</li> +<li>Templates: 140739</li> +<li>Genuine Matches: 7819362</li> +<li>Impostor Matches: 39584639</li> +</ul> +<p>Why not include US Soliders instead of activists?</p> +<p>was creted by Nobilis, a United States Government contractor is used to develop software for the US intelligence agencies as part of the IARPA Janus program.</p> +<p>The IARPA Janus program is</p> +<p>these representations must address the challenges of Aging, Pose, Illumination, and Expression (A-PIE) by exploiting all available imagery.</p> +<ul> +<li>metadata annotations were created using crowd annotations</li> +<li>created by Nobilis</li> +<li>used mechanical turk</li> +<li>made for intelligence analysts</li> +<li>improve performance of face recognition tools</li> +<li>by fusing the rich spatial, temporal, and contextual information available from the multiple views captured by today’s "media in the wild"</li> +</ul> +<p>The name list includes</p> +<ul> +<li>2 videos from CCC<ul> +<li>yq6ZC-YLHZA.png<ul> +<li>Katharina Nocun: Deine Rechte sind in diesen Freihandelsabkommen nicht verfügbar</li> +</ul> +</li> +<li>fF2MxkDzlVg<ul> +<li>Jillian York: "Technology companies now hold an unprecedented ability to shape the world around us by limiting our ability to access certain content and by crafting proprietary algorithm that bring us our daily streams of content. Matthew Stender, Jillian C. York"</li> +</ul> +</li> +<li>Maya Zankoul. She's an old friend, a Lebanese web designer who's put out a couple of books locally and has a Wikipedia page, probably created by a Lebanese Wikipedia editor die-hard. Not famous. How on earth?</li> +<li>Melissa Gira Grant (also a journalist)</li> +<li>Nadezhda Tolokinnikova (Pussy Riot)</li> +<li>Derrick Ashong (activist and journalist)</li> +<li>Michael Anti</li> +<li>Lina Ben Mhenni</li> +<li>Manal al-Sharif</li> +<li>Juan Carlos de Martin (not an activist but not really famous either!)</li> +<li>Anita Sarkeesian</li> +<li>Amal Clooney (lawyer)</li> +<li>Anil Dash (startup guy)</li> +<li>Bruno Latour (philosopher)</li> +<li>Dan Gillmor (tech journalist)</li> +<li>Eben Upton (founder of raspberry pi)</li> +<li>Evgeny Morozov</li> +<li>Gabriella Coleman</li> +<li>Maria Popova</li> +<li>Molly Crabapple</li> +<li>Paola Antonelli</li> +<li>Seymour Hersh</li> +<li>Ta-Nehisi Coates</li> +</ul> +</li> +</ul> +<p>The first 777 are non-alphabetical. From 777-3531 is alphabetical</p> </section><section class='images'><div class='image'><img src='https://nyc3.digitaloceanspaces.com/megapixels/v1/datasets/ijb_c/assets/ijb_c_montage.jpg' alt=' A visualization of the IJB-C dataset'><div class='caption'> A visualization of the IJB-C dataset</div></div></section><section><h2>Research notes</h2> <p>From original papers: <a href="https://noblis.org/wp-content/uploads/2018/03/icb2018.pdf">https://noblis.org/wp-content/uploads/2018/03/icb2018.pdf</a></p> <p>Collection for the dataset began by identifying CreativeCommons subject videos, which are often more scarce thanCreative Commons subject images. Search terms that re-sulted in large quantities of person-centric videos (e.g. “in-terview”) were generated and translated into numerous lan-guages including Arabic, Korean, Swahili, and Hindi to in-crease diversity of the subject pool. Certain YouTube userswho upload well-labeled, person-centric videos, such as the World Economic Forum and the International University Sports Federation were also identified. Titles of videos per-taining to these search terms and usernames were scrapedusing the YouTube Data API and translated into English us-ing the Yandex Translate API4. Pattern matching was per-formed to extract potential names of subjects from the trans-lated titles, and these names were searched using the Wiki-data API to verify the subject’s existence and status as a public figure, and to check for Wikimedia Commons im-agery. Age, gender, and geographic region were collectedusing the Wikipedia API.Using the candidate subject names, Creative Commonsimages were scraped from Google and Wikimedia Com-mons, and Creative Commons videos were scraped fromYouTube. After images and videos of the candidate subjectwere identified, AMT Workers were tasked with validat-ing the subject’s presence throughout the video. The AMTWorkers marked segments of the video in which the subjectwas present, and key frames</p> |