{"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/learning-generalizable-and-identity","title":"Learning Generalizable and Identity-Discriminative Representations for Face Anti-Spoofing","arxiv_id":"1901.05602","date":"2019-01-17","proceeding":null,"authors":["Xiaoguang Tu","Jian Zhao","Mei Xie","Guodong Du","Hengsheng Zhang","Jianshu Li","Zheng Ma","Jiashi Feng"],"abstract":"Face anti-spoofing (a.k.a presentation attack detection) has drawn growing\nattention due to the high-security demand in face authentication systems.\nExisting CNN-based approaches usually well recognize the spoofing faces when\ntraining and testing spoofing samples display similar patterns, but their\nperformance would drop drastically on testing spoofing faces of unseen scenes.\nIn this paper, we try to boost the generalizability and applicability of these\nmethods by designing a CNN model with two major novelties. First, we propose a\nsimple yet effective Total Pairwise Confusion (TPC) loss for CNN training,\nwhich enhances the generalizability of the learned Presentation Attack (PA)\nrepresentations. Secondly, we incorporate a Fast Domain Adaptation (FDA)\ncomponent into the CNN model to alleviate negative effects brought by domain\nchanges. Besides, our proposed model, which is named Generalizable Face\nAuthentication CNN (GFA-CNN), works in a multi-task manner, performing face\nanti-spoofing and face recognition simultaneously. Experimental results show\nthat GFA-CNN outperforms previous face anti-spoofing approaches and also well\npreserves the identity information of input face images.","url_abs":"http://arxiv.org/abs/1901.05602v1","url_pdf":"http://arxiv.org/pdf/1901.05602v1.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":"learning-generalizable-and-identity","repo_url":"https://github.com/tungdt92/GFA-CNN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"domain-adaptation","task_name":"Domain Adaptation"},{"task_slug":"face-anti-spoofing","task_name":"Face Anti-Spoofing"},{"task_slug":"face-recognition","task_name":"Face Recognition"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/face-anti-spoofing-on-msu-mfsd","task":"Face Anti-Spoofing","dataset":"MSU-MFSD","model":"GFA-CNN","rank_in_archive_order":2,"of":3,"metrics":{"Equal Error Rate":"7.5%"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1901.05602","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}