{"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/dpatch-an-adversarial-patch-attack-on-object","title":"DPatch: An Adversarial Patch Attack on Object Detectors","arxiv_id":"1806.02299","date":"2018-06-05","proceeding":null,"authors":["Xin Liu","Huanrui Yang","Ziwei Liu","Linghao Song","Hai Li","Yiran Chen"],"abstract":"Object detectors have emerged as an indispensable module in modern computer\nvision systems. In this work, we propose DPatch -- a black-box\nadversarial-patch-based attack towards mainstream object detectors (i.e. Faster\nR-CNN and YOLO). Unlike the original adversarial patch that only manipulates\nimage-level classifier, our DPatch simultaneously attacks the bounding box\nregression and object classification so as to disable their predictions.\nCompared to prior works, DPatch has several appealing properties: (1) DPatch\ncan perform both untargeted and targeted effective attacks, degrading the mAP\nof Faster R-CNN and YOLO from 75.10% and 65.7% down to below 1%, respectively.\n(2) DPatch is small in size and its attacking effect is location-independent,\nmaking it very practical to implement real-world attacks. (3) DPatch\ndemonstrates great transferability among different detectors as well as\ntraining datasets. For example, DPatch that is trained on Faster R-CNN can\neffectively attack YOLO, and vice versa. Extensive evaluations imply that\nDPatch can perform effective attacks under black-box setup, i.e., even without\nthe knowledge of the attacked network's architectures and parameters.\nSuccessful realization of DPatch also illustrates the intrinsic vulnerability\nof the modern detector architectures to such patch-based adversarial attacks.","url_abs":"http://arxiv.org/abs/1806.02299v4","url_pdf":"http://arxiv.org/pdf/1806.02299v4.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":"dpatch-an-adversarial-patch-attack-on-object","repo_url":"https://github.com/veralauee/DPatch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"object","task_name":"Object"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"faster-r-cnn","method_name":"Faster R-CNN"},{"method_slug":"rpn","method_name":"RPN"},{"method_slug":"roipool","method_name":"RoIPool"},{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1806.02299","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}