{"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/domain-adaptive-self-supervised-pre-training","title":"Domain-Adaptive Self-Supervised Pre-Training for Face & Body Detection in Drawings","arxiv_id":"2211.10641","date":"2022-11-19","proceeding":null,"authors":["Barış Batuhan Topal","Deniz Yuret","Tevfik Metin Sezgin"],"abstract":"Drawings are powerful means of pictorial abstraction and communication. Understanding diverse forms of drawings, including digital arts, cartoons, and comics, has been a major problem of interest for the computer vision and computer graphics communities. Although there are large amounts of digitized drawings from comic books and cartoons, they contain vast stylistic variations, which necessitate expensive manual labeling for training domain-specific recognizers. In this work, we show how self-supervised learning, based on a teacher-student network with a modified student network update design, can be used to build face and body detectors. Our setup allows exploiting large amounts of unlabeled data from the target domain when labels are provided for only a small subset of it. We further demonstrate that style transfer can be incorporated into our learning pipeline to bootstrap detectors using a vast amount of out-of-domain labeled images from natural images (i.e., images from the real world). Our combined architecture yields detectors with state-of-the-art (SOTA) and near-SOTA performance using minimal annotation effort. Our code can be accessed from https://github.com/barisbatuhan/DASS_Detector.","url_abs":"https://arxiv.org/abs/2211.10641v2","url_pdf":"https://arxiv.org/pdf/2211.10641v2.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":"domain-adaptive-self-supervised-pre-training","repo_url":"https://github.com/barisbatuhan/dass_det_inference","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"domain-adaptive-self-supervised-pre-training","repo_url":"https://github.com/barisbatuhan/dass_detector","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"body-detection","task_name":"Body Detection"},{"task_slug":"face-detection","task_name":"Face Detection"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"self-supervised-learning","task_name":"Self-Supervised Learning"},{"task_slug":"style-transfer","task_name":"Style Transfer"},{"task_slug":"weakly-supervised-object-detection","task_name":"Weakly Supervised Object Detection"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/body-detection-on-clipart1k","task":"Body Detection","dataset":"Clipart1k","model":"DASS-Detector (YOLOX XL)","rank_in_archive_order":1,"of":1,"metrics":{"MAP ":"83.59"},"uses_additional_data":false},{"leaderboard":"/sota/body-detection-on-comic2k","task":"Body Detection","dataset":"Comic2k","model":"DASS-Detector (YOLOX XL)","rank_in_archive_order":1,"of":1,"metrics":{"MAP ":"73.65"},"uses_additional_data":false},{"leaderboard":"/sota/body-detection-on-dcm","task":"Body Detection","dataset":"DCM","model":"DASS-Detector (YOLOX XL)","rank_in_archive_order":1,"of":2,"metrics":{"Average Precision":"86.14"},"uses_additional_data":false},{"leaderboard":"/sota/body-detection-on-dcm","task":"Body Detection","dataset":"DCM","model":"DASS-Detector (YOLOX Tiny)","rank_in_archive_order":2,"of":2,"metrics":{"Average Precision":"87.06"},"uses_additional_data":false},{"leaderboard":"/sota/body-detection-on-manga109","task":"Body Detection","dataset":"Manga109","model":"DASS-Detector (YOLOX XL)","rank_in_archive_order":1,"of":1,"metrics":{"Average Precision":"87.98"},"uses_additional_data":false},{"leaderboard":"/sota/body-detection-on-watercolor2k","task":"Body Detection","dataset":"Watercolor2k","model":"DASS-Detector (YOLOX XL)","rank_in_archive_order":1,"of":1,"metrics":{"MAP ":"89.81"},"uses_additional_data":false},{"leaderboard":"/sota/face-detection-on-dcm","task":"Face Detection","dataset":"DCM","model":"DASS-Detector (YOLOX XL)","rank_in_archive_order":1,"of":2,"metrics":{"Average Precision":"77.40"},"uses_additional_data":false},{"leaderboard":"/sota/face-detection-on-dcm","task":"Face Detection","dataset":"DCM","model":"DASS-Detector (YOLOX Tiny)","rank_in_archive_order":2,"of":2,"metrics":{"Average Precision":"77.40"},"uses_additional_data":false},{"leaderboard":"/sota/face-detection-on-manga109","task":"Face Detection","dataset":"Manga109","model":"DASS-Detector (YOLOX XL)","rank_in_archive_order":1,"of":1,"metrics":{"Average Precision":"87.88"},"uses_additional_data":false},{"leaderboard":"/sota/face-detection-on-icartoonface","task":"Face Detection","dataset":"iCartoonFace","model":"DASS-Detector (YOLOX XL)","rank_in_archive_order":1,"of":2,"metrics":{"Average Precision":"90.01"},"uses_additional_data":false},{"leaderboard":"/sota/face-detection-on-icartoonface","task":"Face Detection","dataset":"iCartoonFace","model":"DASS-Detector (YOLOX Tiny)","rank_in_archive_order":2,"of":2,"metrics":{"Average Precision":"87.75"},"uses_additional_data":false},{"leaderboard":"/sota/object-detection-on-manga109","task":"Object Detection","dataset":"Manga109","model":"DASS-Detector (YOLOX XL)","rank_in_archive_order":1,"of":2,"metrics":{"Average Precision":"87.93"},"uses_additional_data":false},{"leaderboard":"/sota/object-detection-on-manga109","task":"Object Detection","dataset":"Manga109","model":"DASS-Detector (YOLOX Tiny)","rank_in_archive_order":2,"of":2,"metrics":{"Average Precision":"87.46"},"uses_additional_data":false},{"leaderboard":"/sota/weakly-supervised-object-detection-on-2","task":"Weakly Supervised Object Detection","dataset":"Clipart1k","model":"DASS-Detector (YOLOX Tiny)","rank_in_archive_order":2,"of":7,"metrics":{"MAP":"64.25"},"uses_additional_data":false},{"leaderboard":"/sota/weakly-supervised-object-detection-on-comic2k","task":"Weakly Supervised Object Detection","dataset":"Comic2k","model":"DASS-Detector (YOLOX Tiny)","rank_in_archive_order":1,"of":8,"metrics":{"MAP":"67.41"},"uses_additional_data":true},{"leaderboard":"/sota/weakly-supervised-object-detection-on-1","task":"Weakly Supervised Object Detection","dataset":"Watercolor2k","model":"DASS-Detector (YOLOX Tiny)","rank_in_archive_order":2,"of":12,"metrics":{"MAP":"71.53"},"uses_additional_data":true}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2211.10641","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}