{"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/learning-intelligent-dialogs-for-bounding-box","title":"Learning Intelligent Dialogs for Bounding Box Annotation","arxiv_id":"1712.08087","date":"2017-12-21","proceeding":"CVPR 2018 6","authors":["Ksenia Konyushkova","Jasper Uijlings","Christoph Lampert","Vittorio Ferrari"],"abstract":"We introduce Intelligent Annotation Dialogs for bounding box annotation. We\ntrain an agent to automatically choose a sequence of actions for a human\nannotator to produce a bounding box in a minimal amount of time. Specifically,\nwe consider two actions: box verification, where the annotator verifies a box\ngenerated by an object detector, and manual box drawing. We explore two kinds\nof agents, one based on predicting the probability that a box will be\npositively verified, and the other based on reinforcement learning. We\ndemonstrate that (1) our agents are able to learn efficient annotation\nstrategies in several scenarios, automatically adapting to the image\ndifficulty, the desired quality of the boxes, and the detector strength; (2) in\nall scenarios the resulting annotation dialogs speed up annotation compared to\nmanual box drawing alone and box verification alone, while also outperforming\nany fixed combination of verification and drawing in most scenarios; (3) in a\nrealistic scenario where the detector is iteratively re-trained, our agents\nevolve a series of strategies that reflect the shifting trade-off between\nverification and drawing as the detector grows stronger.","url_abs":"http://arxiv.org/abs/1712.08087v3","url_pdf":"http://arxiv.org/pdf/1712.08087v3.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":"learning-intelligent-dialogs-for-bounding-box","repo_url":"https://github.com/google/intelligent_annotation_dialogs","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"}],"methods":[{"method_slug":"speed","method_name":"SPEED"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1712.08087","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}