{"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/fast-geometrically-perturbed-adversarial","title":"Fast Geometrically-Perturbed Adversarial Faces","arxiv_id":"1809.08999","date":"2018-09-24","proceeding":null,"authors":["Ali Dabouei","Sobhan Soleymani","Jeremy Dawson","Nasser M. Nasrabadi"],"abstract":"The state-of-the-art performance of deep learning algorithms has led to a\nconsiderable increase in the utilization of machine learning in\nsecurity-sensitive and critical applications. However, it has recently been\nshown that a small and carefully crafted perturbation in the input space can\ncompletely fool a deep model. In this study, we explore the extent to which\nface recognition systems are vulnerable to geometrically-perturbed adversarial\nfaces. We propose a fast landmark manipulation method for generating\nadversarial faces, which is approximately 200 times faster than the previous\ngeometric attacks and obtains 99.86% success rate on the state-of-the-art face\nrecognition models. To further force the generated samples to be natural, we\nintroduce a second attack constrained on the semantic structure of the face\nwhich has the half speed of the first attack with the success rate of 99.96%.\nBoth attacks are extremely robust against the state-of-the-art defense methods\nwith the success rate of equal or greater than 53.59%. Code is available at\nhttps://github.com/alldbi/FLM","url_abs":"http://arxiv.org/abs/1809.08999v2","url_pdf":"http://arxiv.org/pdf/1809.08999v2.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":"fast-geometrically-perturbed-adversarial","repo_url":"https://github.com/alldbi/FLM","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"face-recognition","task_name":"Face Recognition"}],"methods":[{"method_slug":"speed","method_name":"SPEED"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1809.08999","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}