{"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/person-re-identification-with-deep-similarity","title":"Person Re-identification with Deep Similarity-Guided Graph Neural Network","arxiv_id":"1807.09975","date":"2018-07-26","proceeding":"ECCV 2018 9","authors":["Yantao Shen","Hongsheng Li","Shuai Yi","Dapeng Chen","Xiaogang Wang"],"abstract":"The person re-identification task requires to robustly estimate visual\nsimilarities between person images. However, existing person re-identification\nmodels mostly estimate the similarities of different image pairs of probe and\ngallery images independently while ignores the relationship information between\ndifferent probe-gallery pairs. As a result, the similarity estimation of some\nhard samples might not be accurate. In this paper, we propose a novel deep\nlearning framework, named Similarity-Guided Graph Neural Network (SGGNN) to\novercome such limitations. Given a probe image and several gallery images,\nSGGNN creates a graph to represent the pairwise relationships between\nprobe-gallery pairs (nodes) and utilizes such relationships to update the\nprobe-gallery relation features in an end-to-end manner. Accurate similarity\nestimation can be achieved by using such updated probe-gallery relation\nfeatures for prediction. The input features for nodes on the graph are the\nrelation features of different probe-gallery image pairs. The probe-gallery\nrelation feature updating is then performed by the messages passing in SGGNN,\nwhich takes other nodes' information into account for similarity estimation.\nDifferent from conventional GNN approaches, SGGNN learns the edge weights with\nrich labels of gallery instance pairs directly, which provides relation fusion\nmore precise information. The effectiveness of our proposed method is validated\non three public person re-identification datasets.","url_abs":"http://arxiv.org/abs/1807.09975v1","url_pdf":"http://arxiv.org/pdf/1807.09975v1.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":"graph-neural-network","task_name":"Graph Neural Network"},{"task_slug":"person-re-identification","task_name":"Person Re-Identification"},{"task_slug":null,"task_name":"Relation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/person-re-identification-on-cuhk03","task":"Person Re-Identification","dataset":"CUHK03","model":"Proposed SGGNN","rank_in_archive_order":1,"of":19,"metrics":{"MAP":"94.3"},"uses_additional_data":false},{"leaderboard":"/sota/person-re-identification-on-dukemtmc-reid","task":"Person Re-Identification","dataset":"DukeMTMC-reID","model":"Proposed SGGNN","rank_in_archive_order":67,"of":94,"metrics":{"mAP":"68.2"},"uses_additional_data":false},{"leaderboard":"/sota/person-re-identification-on-market-1501","task":"Person Re-Identification","dataset":"Market-1501","model":"Proposed SGGNN","rank_in_archive_order":133,"of":135,"metrics":{"mAP":"82.8"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1807.09975","atlas_url":"https://app.syntology.ai/?focus=1807.09975","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}