[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-187866-en":3,"doc-seo-187866-105":30,"detail-sidebar-cat-1-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":11,"category_id":12,"category_name":13,"doc_title":14,"doc_description":15,"doc_content":16,"file_id":17,"file_url":18,"file_type":19,"file_size":20,"view_count":11,"is_deleted":4,"is_public":11,"is_downloadable":11,"audit_status":11,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":15,"update_tm":28,"read_time":29},187866,19241457091524,"Gelato","https://us-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",1,21,"Paper Templates","Human Re-identification by Matching Compositional Template with Cluster Sampling","This paper addresses human re-identification at a distance by matching body information using several reference examples. It targets key challenges caused by large human appearance variability and high false positives from pose, illumination, occlusion, and clutter. The method builds an expressive compositional template from a few reference images, representing the body as an articulated assembly of compositional and alternative parts. An effective cluster-sampling matching algorithm operates on a candidacy graph with two-step convergence, and improves performance on three public datasets.","# Human Re-identification by Matching Compositional Template withCluster Sampling\n\nYuanlu Xu¹Liang Lin¹*  Wei-Shi Zheng¹Xiaobai Liu²¹Sun Yat-Sen University,China2University of California,Los Angelesmerayxu@gmail.com,{1inliang,wszheng}@ieee.org,lxb@ucla.edu  \n## Abstract\n\nThis paper aims at a newly raising task in visual surveil-lance:re-identifying people at a distance by matching bodyinformation,given several reference examples.Most of ex-isting works solve this task by matching a reference tem-plate with the target individual,but often suffer from largehuman appearance variability(e.g.different poses/views,illumination)and high false positives in matching causedbyconjunctions,occlusions or surrounding clutters.Address-ing these problems,we construct a simple yet expressivetemplate from a few reference images of a certain individ-ual,which represents the body as an articulated assembly ofcompositional and alternative parts,and propose an effec-tive matching algorithm with cluster sampling.This algo-rithm is designed within a candidacy graph whose verticesare matching candidates (i.e.a pair of source and targetbody parts),and iterates in two steps for convergence.(i)It generates possible partial matches based on compatibleand competitive relations among body parts.(ii)It con-firms the partial matches to generate a new matching solu-tion,which is accepted by the Markov Chain Monte Carlo(MCMC)mechanism.In the experiments,we demonstratethe superior performance of our approach on three publicdatabases compared to existing methods.  \n## 1.Introduction\n\nPerson re-identification at a distance increasingly re-ceives attention in video surveillance,particularly for theapplications restricting the use of face recognition.But thistask is very challenging due to the following difficulties,  \n● Robust human representation(signature).There arelarge variations for human body in appearance,(e.g.,dif-ferent views,poses,lighting conditions).It is usually in-  \nFigure 1.An illustration of the proposed approach.A query indi-vidual is represented as a compositional part-based template,andpart proposals are extracted from multiple instances at each parts.Human re-identification is thus posed as compositional templatematching.Note certain parts are omitted for clear specification.  \ntractable to construct a template of the individual to be rec-  \nognized by extracting only low-level image features.  \n● Effective human matching (localizing).Given thetemplate,re-identifying targets with the global body infor-mation often suffers from high matching false positives,asthe targets are possibly occluded or conjuncted with othersand backgrounds in realistic surveillance applications.Fur-thermore,it is desired to accurately localize human bodyparts in general.  \nThe objective of human re-identification in this work isto recognize an individual by employing body informationto address the above difficulties.We study the problem withthe following setting based on the application requirementsin surveillance:(1)The clothing of individuals remain un-changed across different scenarios.(2)The individual tobe re-identified should be in a moderate resolution,(e.g.,  \n≥120 pixels in height).Our approach builds a compo-sitional part-based template to represent the target individ-ual and matches the template with input images by employ-ing a stochastic cluster sampling algorithm,as illustrated inFig.1.  \nWe organize the template of a query individual with anexpressive tree representation that can be produced in avery simple way.We perform the human body part de-tectors [1,2]on several reference images of the individ-ual,and the images of detected parts are grouped accord-ing to their semantics.That is,a human template is de-composed into body parts,e.g.,head,torso,arms,each ofwhich associates with a number of part instances.Notethat we can prune the instances sharing very similar ap-pearances with others.This expressive template fully ex-ploit information from multiple r","cbCaii0lFfXcZWii","https://ap.wps.com/l/cbCaii0lFfXcZWii","pdf",1464295,8,"English","en",105,"# Abstract\n# Introduction","[{\"question\":\"What problem does the paper focus on?\",\"answer\":\"The paper focuses on re-identifying people at a distance in visual surveillance by matching body information using reference examples.\"},{\"question\":\"What main challenges does the approach aim to solve?\",\"answer\":\"It targets robustness against appearance variability (different views, poses, and illumination) and reduces matching false positives caused by occlusions, conjunctions, and surrounding clutter.\"},{\"question\":\"How does the proposed method perform matching?\",\"answer\":\"It constructs an expressive compositional template from multiple reference images and uses a stochastic cluster-sampling algorithm on a candidacy graph, iterating in two steps to generate and confirm partial matches via Markov Chain Monte Carlo.\"}]","Human Re-identification by Matching Compositional Template with Cluster Sampling | PDF",1788385840,3,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":14,"keywords":34,"description":15,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"human-re-identification-by-matching-compositional-template-with-cluster-sampling","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,50],{"item":41,"name":42,"@type":43,"position":11},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/template/","Template",2,{"item":49,"name":13,"@type":43,"position":29},"https://docshare.wps.com/template/paper-templates/",{"item":51,"name":14,"@type":43,"position":52},"https://docshare.wps.com/template/human-re-identification-by-matching-compositional-template-with-cluster-sampling/187866/",4,{"url":51,"name":14,"@type":54,"author":55,"headline":14,"publisher":57,"fileFormat":60,"inLanguage":23,"description":15,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-09-07","2026-09-02",true,{"@type":65,"interactionType":66,"userInteractionCount":29},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the paper focus on?","Question",{"text":75,"@type":76},"The paper focuses on re-identifying people at a distance in visual surveillance by matching body information using reference examples.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What main challenges does the approach aim to solve?",{"text":80,"@type":76},"It targets robustness against appearance variability (different views, poses, and illumination) and reduces matching false positives caused by occlusions, conjunctions, and surrounding clutter.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the proposed method perform matching?",{"text":84,"@type":76},"It constructs an expressive compositional template from multiple reference images and uses a stochastic cluster-sampling algorithm on a candidacy graph, iterating in two steps to generate and confirm partial matches via Markov Chain Monte Carlo.","https://schema.org",{"og:url":51,"og:type":87,"og:title":14,"og:site_name":58,"og:description":15},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,98,103,108,113,118,123,128,131],{"id":94,"doc_module":11,"doc_module_name":46,"category_name":95,"show_sort_weight":96,"slug":97},11,"Presentations",90,"presentations",{"id":99,"doc_module":11,"doc_module_name":46,"category_name":100,"show_sort_weight":101,"slug":102},12,"Resumes",80,"resumes",{"id":104,"doc_module":11,"doc_module_name":46,"category_name":105,"show_sort_weight":106,"slug":107},14,"Invoices",70,"invoices",{"id":109,"doc_module":11,"doc_module_name":46,"category_name":110,"show_sort_weight":111,"slug":112},15,"Posters",60,"posters",{"id":114,"doc_module":11,"doc_module_name":46,"category_name":115,"show_sort_weight":116,"slug":117},16,"Social Media",50,"social-media",{"id":119,"doc_module":11,"doc_module_name":46,"category_name":120,"show_sort_weight":121,"slug":122},17,"Forms",40,"forms",{"id":124,"doc_module":11,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},18,"Letters",30,"letters",{"id":12,"doc_module":11,"doc_module_name":46,"category_name":13,"show_sort_weight":129,"slug":130},5,"papers-templates",{"id":132,"doc_module":11,"doc_module_name":46,"category_name":133,"show_sort_weight":4,"slug":134},158,"General","general-158"]