{"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/joint-detection-and-identification-feature","title":"Joint Detection and Identification Feature Learning for Person Search","arxiv_id":"1604.01850","date":"2016-04-07","proceeding":"CVPR 2017 7","authors":["Tong Xiao","Shuang Li","Bochao Wang","Liang Lin","Xiaogang Wang"],"abstract":"Existing person re-identification benchmarks and methods mainly focus on\nmatching cropped pedestrian images between queries and candidates. However, it\nis different from real-world scenarios where the annotations of pedestrian\nbounding boxes are unavailable and the target person needs to be searched from\na gallery of whole scene images. To close the gap, we propose a new deep\nlearning framework for person search. Instead of breaking it down into two\nseparate tasks---pedestrian detection and person re-identification, we jointly\nhandle both aspects in a single convolutional neural network. An Online\nInstance Matching (OIM) loss function is proposed to train the network\neffectively, which is scalable to datasets with numerous identities. To\nvalidate our approach, we collect and annotate a large-scale benchmark dataset\nfor person search. It contains 18,184 images, 8,432 identities, and 96,143\npedestrian bounding boxes. Experiments show that our framework outperforms\nother separate approaches, and the proposed OIM loss function converges much\nfaster and better than the conventional Softmax loss.","url_abs":"http://arxiv.org/abs/1604.01850v3","url_pdf":"http://arxiv.org/pdf/1604.01850v3.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":"joint-detection-and-identification-feature","repo_url":"https://github.com/ShuangLI59/person_search","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"joint-detection-and-identification-feature","repo_url":"https://github.com/ChrisLee63/person_search","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"pedestrian-detection","task_name":"Pedestrian Detection"},{"task_slug":"person-re-identification","task_name":"Person Re-Identification"},{"task_slug":"person-search","task_name":"Person Search"}],"methods":[{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[{"slug":"cuhk-sysu","name":"CUHK-SYSU","full_name":"CUHK-SYSU Person Search Dataset"}],"methods_introduced":[],"results":[{"leaderboard":"/sota/person-re-identification-on-cuhk03","task":"Person Re-Identification","dataset":"CUHK03","model":"OIM Loss 45","rank_in_archive_order":10,"of":19,"metrics":{"MAP":"72.5","Rank-1":"77.5"},"uses_additional_data":false},{"leaderboard":"/sota/person-re-identification-on-dukemtmc-reid","task":"Person Re-Identification","dataset":"DukeMTMC-reID","model":"OIM","rank_in_archive_order":85,"of":94,"metrics":{"Rank-1":"68.1","mAP":"47.4"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1604.01850","atlas_url":"https://app.syntology.ai/?focus=1604.01850","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}