{"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-representation-learning-with-part-loss","title":"Deep Representation Learning with Part Loss for Person Re-Identification","arxiv_id":"1707.00798","date":"2017-07-04","proceeding":null,"authors":["Hantao Yao","Shiliang Zhang","Yongdong Zhang","Jintao Li","Qi Tian"],"abstract":"Learning discriminative representations for unseen person images is critical\nfor person Re-Identification (ReID). Most of current approaches learn deep\nrepresentations in classification tasks, which essentially minimize the\nempirical classification risk on the training set. As shown in our experiments,\nsuch representations commonly focus on several body parts discriminative to the\ntraining set, rather than the entire human body. Inspired by the structural\nrisk minimization principle in SVM, we revise the traditional deep\nrepresentation learning procedure to minimize both the empirical classification\nrisk and the representation learning risk. The representation learning risk is\nevaluated by the proposed part loss, which automatically generates several\nparts for an image, and computes the person classification loss on each part\nseparately. Compared with traditional global classification loss,\nsimultaneously considering multiple part loss enforces the deep network to\nfocus on the entire human body and learn discriminative representations for\ndifferent parts. Experimental results on three datasets, i.e., Market1501,\nCUHK03, VIPeR, show that our representation outperforms the existing deep\nrepresentations.","url_abs":"http://arxiv.org/abs/1707.00798v2","url_pdf":"http://arxiv.org/pdf/1707.00798v2.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":"classification-1","task_name":"Classification"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"person-re-identification","task_name":"Person Re-Identification"},{"task_slug":"representation-learning","task_name":"Representation Learning"}],"methods":[{"method_slug":"svm","method_name":"SVM"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/person-re-identification-on-market-1501","task":"Person Re-Identification","dataset":"Market-1501","model":"PartLoss","rank_in_archive_order":99,"of":135,"metrics":{"Rank-1":"88.2","mAP":"69.3"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1707.00798","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}