{"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/we-dont-need-no-bounding-boxes-training","title":"We don't need no bounding-boxes: Training object class detectors using only human verification","arxiv_id":"1602.08405","date":"2016-02-26","proceeding":"CVPR 2016 6","authors":["Dim P. Papadopoulos","Jasper R. R. Uijlings","Frank Keller","Vittorio Ferrari"],"abstract":"Training object class detectors typically requires a large set of images in\nwhich objects are annotated by bounding-boxes. However, manually drawing\nbounding-boxes is very time consuming. We propose a new scheme for training\nobject detectors which only requires annotators to verify bounding-boxes\nproduced automatically by the learning algorithm. Our scheme iterates between\nre-training the detector, re-localizing objects in the training images, and\nhuman verification. We use the verification signal both to improve re-training\nand to reduce the search space for re-localisation, which makes these steps\ndifferent to what is normally done in a weakly supervised setting. Extensive\nexperiments on PASCAL VOC 2007 show that (1) using human verification to update\ndetectors and reduce the search space leads to the rapid production of\nhigh-quality bounding-box annotations; (2) our scheme delivers detectors\nperforming almost as good as those trained in a fully supervised setting,\nwithout ever drawing any bounding-box; (3) as the verification task is very\nquick, our scheme substantially reduces total annotation time by a factor\n6x-9x.","url_abs":"http://arxiv.org/abs/1602.08405v3","url_pdf":"http://arxiv.org/pdf/1602.08405v3.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":"we-dont-need-no-bounding-boxes-training","repo_url":"https://github.com/EscVM/OIDv4_ToolKit","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1602.08405","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}