{"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-co-attention-based-comparators-for","title":"Deep Co-attention based Comparators For Relative Representation Learning in Person Re-identification","arxiv_id":"1804.11027","date":"2018-04-30","proceeding":null,"authors":["Lin Wu","Yang Wang","Junbin Gao","DaCheng Tao"],"abstract":"Person re-identification (re-ID) requires rapid, flexible yet discriminant\nrepresentations to quickly generalize to unseen observations on-the-fly and\nrecognize the same identity across disjoint camera views. Recent effective\nmethods are developed in a pair-wise similarity learning system to detect a\nfixed set of features from distinct regions which are mapped to their vector\nembeddings for the distance measuring. However, the most relevant and crucial\nparts of each image are detected independently without referring to the\ndependency conditioned on one and another. Also, these region based methods\nrely on spatial manipulation to position the local features in comparable\nsimilarity measuring. To combat these limitations, in this paper we introduce\nthe Deep Co-attention based Comparators (DCCs) that fuse the co-dependent\nrepresentations of the paired images so as to focus on the relevant parts of\nboth images and produce their \\textit{relative representations}. Given a pair\nof pedestrian images to be compared, the proposed model mimics the foveation of\nhuman eyes to detect distinct regions concurrent on both images, namely\nco-dependent features, and alternatively attend to relevant regions to fuse\nthem into the similarity learning. Our comparator is capable of producing\ndynamic representations relative to a particular sample every time, and thus\nwell-suited to the case of re-identifying pedestrians on-the-fly. We perform\nextensive experiments to provide the insights and demonstrate the effectiveness\nof the proposed DCCs in person re-ID. Moreover, our approach has achieved the\nstate-of-the-art performance on three benchmark data sets: DukeMTMC-reID\n\\cite{DukeMTMC}, CUHK03 \\cite{FPNN}, and Market-1501 \\cite{Market1501}.","url_abs":"http://arxiv.org/abs/1804.11027v1","url_pdf":"http://arxiv.org/pdf/1804.11027v1.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":"deep-co-attention-based-comparators-for","repo_url":"https://github.com/gitabcworld/ConvArc","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"foveation","task_name":"Foveation"},{"task_slug":"person-re-identification","task_name":"Person Re-Identification"},{"task_slug":"representation-learning","task_name":"Representation Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}