{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/consensus-driven-propagation-in-massive","title":"Consensus-Driven Propagation in Massive Unlabeled Data for Face Recognition","arxiv_id":"1809.01407","date":"2018-09-05","proceeding":"ECCV 2018 9","authors":["Xiaohang Zhan","Ziwei Liu","Junjie Yan","Dahua Lin","Chen Change Loy"],"abstract":"Face recognition has witnessed great progress in recent years, mainly\nattributed to the high-capacity model designed and the abundant labeled data\ncollected. However, it becomes more and more prohibitive to scale up the\ncurrent million-level identity annotations. In this work, we show that\nunlabeled face data can be as effective as the labeled ones. Here, we consider\na setting closely mimicking the real-world scenario, where the unlabeled data\nare collected from unconstrained environments and their identities are\nexclusive from the labeled ones. Our main insight is that although the class\ninformation is not available, we can still faithfully approximate these\nsemantic relationships by constructing a relational graph in a bottom-up\nmanner. We propose Consensus-Driven Propagation (CDP) to tackle this\nchallenging problem with two modules, the \"committee\" and the \"mediator\", which\nselect positive face pairs robustly by carefully aggregating multi-view\ninformation. Extensive experiments validate the effectiveness of both modules\nto discard outliers and mine hard positives. With CDP, we achieve a compelling\naccuracy of 78.18% on MegaFace identification challenge by using only 9% of the\nlabels, comparing to 61.78% when no unlabeled data are used and 78.52% when all\nlabels are employed.","url_abs":"http://arxiv.org/abs/1809.01407v2","url_pdf":"http://arxiv.org/pdf/1809.01407v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"consensus-driven-propagation-in-massive","repo_url":"https://github.com/XiaohangZhan/cdp","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":null},{"paper_slug":"consensus-driven-propagation-in-massive","repo_url":"https://github.com/2023-MindSpore-4/Code8/tree/main/cdp","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"consensus-driven-propagation-in-massive","repo_url":"https://github.com/2024-MindSpore-1/Code7/tree/main/cdp","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"consensus-driven-propagation-in-massive","repo_url":"https://github.com/MindSpore-paper-code-3/code6/tree/main/cdp","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"consensus-driven-propagation-in-massive","repo_url":"https://github.com/Recognito-Vision/Android-FaceRecognition-FaceLivenessDetection","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"face-recognition","task_name":"Face Recognition"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1809.01407","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}