{"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/on-face-segmentation-face-swapping-and-face","title":"On Face Segmentation, Face Swapping, and Face Perception","arxiv_id":"1704.06729","date":"2017-04-22","proceeding":null,"authors":["Yuval Nirkin","Iacopo Masi","Anh Tuan Tran","Tal Hassner","Gerard Medioni"],"abstract":"We show that even when face images are unconstrained and arbitrarily paired,\nface swapping between them is actually quite simple. To this end, we make the\nfollowing contributions. (a) Instead of tailoring systems for face\nsegmentation, as others previously proposed, we show that a standard fully\nconvolutional network (FCN) can achieve remarkably fast and accurate\nsegmentations, provided that it is trained on a rich enough example set. For\nthis purpose, we describe novel data collection and generation routines which\nprovide challenging segmented face examples. (b) We use our segmentations to\nenable robust face swapping under unprecedented conditions. (c) Unlike previous\nwork, our swapping is robust enough to allow for extensive quantitative tests.\nTo this end, we use the Labeled Faces in the Wild (LFW) benchmark and measure\nthe effect of intra- and inter-subject face swapping on recognition. We show\nthat our intra-subject swapped faces remain as recognizable as their sources,\ntestifying to the effectiveness of our method. In line with well known\nperceptual studies, we show that better face swapping produces less\nrecognizable inter-subject results. This is the first time this effect was\nquantitatively demonstrated for machine vision systems.","url_abs":"http://arxiv.org/abs/1704.06729v1","url_pdf":"http://arxiv.org/pdf/1704.06729v1.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":"on-face-segmentation-face-swapping-and-face","repo_url":"https://github.com/YuvalNirkin/face_video_segment","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}},{"paper_slug":"on-face-segmentation-face-swapping-and-face","repo_url":"https://github.com/yuvalnirkin/face_segmentation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"face-swapping","task_name":"Face Swapping"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1704.06729","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}