{"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/deep-self-taught-learning-for-weakly","title":"Deep Self-Taught Learning for Weakly Supervised Object Localization","arxiv_id":"1704.05188","date":"2017-04-18","proceeding":"CVPR 2017 7","authors":["Zequn Jie","Yunchao Wei","Xiaojie Jin","Jiashi Feng","Wei Liu"],"abstract":"Most existing weakly supervised localization (WSL) approaches learn detectors\nby finding positive bounding boxes based on features learned with image-level\nsupervision. However, those features do not contain spatial location related\ninformation and usually provide poor-quality positive samples for training a\ndetector. To overcome this issue, we propose a deep self-taught learning\napproach, which makes the detector learn the object-level features reliable for\nacquiring tight positive samples and afterwards re-train itself based on them.\nConsequently, the detector progressively improves its detection ability and\nlocalizes more informative positive samples. To implement such self-taught\nlearning, we propose a seed sample acquisition method via image-to-object\ntransferring and dense subgraph discovery to find reliable positive samples for\ninitializing the detector. An online supportive sample harvesting scheme is\nfurther proposed to dynamically select the most confident tight positive\nsamples and train the detector in a mutual boosting way. To prevent the\ndetector from being trapped in poor optima due to overfitting, we propose a new\nrelative improvement of predicted CNN scores for guiding the self-taught\nlearning process. Extensive experiments on PASCAL 2007 and 2012 show that our\napproach outperforms the state-of-the-arts, strongly validating its\neffectiveness.","url_abs":"http://arxiv.org/abs/1704.05188v2","url_pdf":"http://arxiv.org/pdf/1704.05188v2.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":[],"tasks":[{"task_slug":"object","task_name":"Object"},{"task_slug":"object-localization","task_name":"Object Localization"},{"task_slug":"weakly-supervised-object-detection","task_name":"Weakly Supervised Object Detection"},{"task_slug":"weakly-supervised-object-localization","task_name":"Weakly-Supervised Object Localization"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/weakly-supervised-object-detection-on-pascal-1","task":"Weakly Supervised Object Detection","dataset":"PASCAL VOC 2007","model":"Deep Self-Taught Learning","rank_in_archive_order":32,"of":41,"metrics":{"MAP":"43.7"},"uses_additional_data":false},{"leaderboard":"/sota/weakly-supervised-object-detection-on-pascal","task":"Weakly Supervised Object Detection","dataset":"PASCAL VOC 2012 test","model":"Deep Self-Taught Learning","rank_in_archive_order":26,"of":32,"metrics":{"MAP":"38.3"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1704.05188","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}