Loading…
Enhanced multi-dataset transfer learning method for unsupervised person re-identification using co-training strategy
This study proposes progressive unsupervised co-learning for unsupervised person re-identification by introducing a co-training strategy in an iterative training process. The authors’ method adopts an iterative training process to improve transferred models by iterating among clustering, selection,...
Saved in:
Published in: | IET computer vision 2018-12, Vol.12 (8), p.1219-1227 |
---|---|
Main Authors: | , |
Format: | Article |
Language: | English |
Subjects: | |
Citations: | Items that this one cites Items that cite this one |
Online Access: | Request full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
cited_by | cdi_FETCH-LOGICAL-c3307-cf02a1d783310a521408e4ffa79346208e36160bba65fedd70c76d17104794c03 |
---|---|
cites | cdi_FETCH-LOGICAL-c3307-cf02a1d783310a521408e4ffa79346208e36160bba65fedd70c76d17104794c03 |
container_end_page | 1227 |
container_issue | 8 |
container_start_page | 1219 |
container_title | IET computer vision |
container_volume | 12 |
creator | Xian, Yuqiao Hu, Haifeng |
description | This study proposes progressive unsupervised co-learning for unsupervised person re-identification by introducing a co-training strategy in an iterative training process. The authors’ method adopts an iterative training process to improve transferred models by iterating among clustering, selection, exchange, and fine-tuning. To solve the problem of transferring representations learned from multiple source datasets, their method utilises multiple convolutional neural network (CNN) models trained on different labelled source datasets by feeding soft labels obtained by clustering on target dataset to each other. The enhanced model can learn more discriminative person representations than the single model trained on multiple datasets. Experimental results on two large-scale benchmark datasets (i.e. DukeMTMC-reID and Market-1501) demonstrate that their method can enhance transferred CNN models by using more source datasets and is competitive to the state-of-the-art methods. |
doi_str_mv | 10.1049/iet-cvi.2018.5103 |
format | article |
fullrecord | <record><control><sourceid>wiley_24P</sourceid><recordid>TN_cdi_doaj_primary_oai_doaj_org_article_d090f21e9c0f49c588d2a7525d40cd40</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><doaj_id>oai_doaj_org_article_d090f21e9c0f49c588d2a7525d40cd40</doaj_id><sourcerecordid>CVI2BF00517</sourcerecordid><originalsourceid>FETCH-LOGICAL-c3307-cf02a1d783310a521408e4ffa79346208e36160bba65fedd70c76d17104794c03</originalsourceid><addsrcrecordid>eNqFkc9u3CAQh62qlZqmfYDe_AJshz82prd2laQrRcol7RURGDasvBABTrVvX5ytcmwPiGHE9400v677TGFDQagvASuxz2HDgE6bgQJ_011QOVCiRgFvX2vO3ncfSjkADKNS4qKrV_HRRIuuPy5zDcSZagrWvmYTi8fcz2hyDHHfH7E-Jtf7lPslluUJ83MojWtFSbHPSILDWIMP1tTQOktZMZtIc4UXRWlVxf3pY_fOm7ngp7_3Zffz-up--4Pc3t3stt9uieUcJLEemKFOTpxTMAOjAiYU3hupuBhZe_CRjvDwYMbBo3MSrBwdlW0hUgkL_LLbnb0umYN-yuFo8kknE_RLI-W9NrkGO6N2oMAzisqCF8oO0-SYkQMbnADbTnPRs8vmVEpG_-qjoNcIdItAtwj0GoFeI2jM1zPzO8x4-j-gt7927Pt1C4fKBpMzvH47pCXHtqp_DPsDXtGeHw</addsrcrecordid><sourcetype>Open Website</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Enhanced multi-dataset transfer learning method for unsupervised person re-identification using co-training strategy</title><source>Wiley Online Library Open Access</source><creator>Xian, Yuqiao ; Hu, Haifeng</creator><creatorcontrib>Xian, Yuqiao ; Hu, Haifeng</creatorcontrib><description>This study proposes progressive unsupervised co-learning for unsupervised person re-identification by introducing a co-training strategy in an iterative training process. The authors’ method adopts an iterative training process to improve transferred models by iterating among clustering, selection, exchange, and fine-tuning. To solve the problem of transferring representations learned from multiple source datasets, their method utilises multiple convolutional neural network (CNN) models trained on different labelled source datasets by feeding soft labels obtained by clustering on target dataset to each other. The enhanced model can learn more discriminative person representations than the single model trained on multiple datasets. Experimental results on two large-scale benchmark datasets (i.e. DukeMTMC-reID and Market-1501) demonstrate that their method can enhance transferred CNN models by using more source datasets and is competitive to the state-of-the-art methods.</description><identifier>ISSN: 1751-9632</identifier><identifier>ISSN: 1751-9640</identifier><identifier>EISSN: 1751-9640</identifier><identifier>DOI: 10.1049/iet-cvi.2018.5103</identifier><language>eng</language><publisher>The Institution of Engineering and Technology</publisher><subject>CNN models ; co-training strategy ; discriminative person representations ; enhanced multidataset transfer learning method ; image recognition ; image representation ; iterative methods ; iterative training process ; labelled source datasets ; large-scale benchmark datasets ; multiple convolutional neural network models ; multiple source datasets ; neural nets ; pattern clustering ; progressive unsupervised co-learning ; Research Article ; single model ; soft labels ; target dataset clustering ; transferred models ; unsupervised learning ; unsupervised person re-identification</subject><ispartof>IET computer vision, 2018-12, Vol.12 (8), p.1219-1227</ispartof><rights>The Institution of Engineering and Technology</rights><rights>2018 The Institution of Engineering and Technology</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c3307-cf02a1d783310a521408e4ffa79346208e36160bba65fedd70c76d17104794c03</citedby><cites>FETCH-LOGICAL-c3307-cf02a1d783310a521408e4ffa79346208e36160bba65fedd70c76d17104794c03</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://onlinelibrary.wiley.com/doi/pdf/10.1049%2Fiet-cvi.2018.5103$$EPDF$$P50$$Gwiley$$H</linktopdf><linktohtml>$$Uhttps://onlinelibrary.wiley.com/doi/full/10.1049%2Fiet-cvi.2018.5103$$EHTML$$P50$$Gwiley$$H</linktohtml><link.rule.ids>314,780,784,9755,11562,27924,27925,46052,46476</link.rule.ids><linktorsrc>$$Uhttps://onlinelibrary.wiley.com/doi/abs/10.1049%2Fiet-cvi.2018.5103$$EView_record_in_Wiley-Blackwell$$FView_record_in_$$GWiley-Blackwell</linktorsrc></links><search><creatorcontrib>Xian, Yuqiao</creatorcontrib><creatorcontrib>Hu, Haifeng</creatorcontrib><title>Enhanced multi-dataset transfer learning method for unsupervised person re-identification using co-training strategy</title><title>IET computer vision</title><description>This study proposes progressive unsupervised co-learning for unsupervised person re-identification by introducing a co-training strategy in an iterative training process. The authors’ method adopts an iterative training process to improve transferred models by iterating among clustering, selection, exchange, and fine-tuning. To solve the problem of transferring representations learned from multiple source datasets, their method utilises multiple convolutional neural network (CNN) models trained on different labelled source datasets by feeding soft labels obtained by clustering on target dataset to each other. The enhanced model can learn more discriminative person representations than the single model trained on multiple datasets. Experimental results on two large-scale benchmark datasets (i.e. DukeMTMC-reID and Market-1501) demonstrate that their method can enhance transferred CNN models by using more source datasets and is competitive to the state-of-the-art methods.</description><subject>CNN models</subject><subject>co-training strategy</subject><subject>discriminative person representations</subject><subject>enhanced multidataset transfer learning method</subject><subject>image recognition</subject><subject>image representation</subject><subject>iterative methods</subject><subject>iterative training process</subject><subject>labelled source datasets</subject><subject>large-scale benchmark datasets</subject><subject>multiple convolutional neural network models</subject><subject>multiple source datasets</subject><subject>neural nets</subject><subject>pattern clustering</subject><subject>progressive unsupervised co-learning</subject><subject>Research Article</subject><subject>single model</subject><subject>soft labels</subject><subject>target dataset clustering</subject><subject>transferred models</subject><subject>unsupervised learning</subject><subject>unsupervised person re-identification</subject><issn>1751-9632</issn><issn>1751-9640</issn><issn>1751-9640</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2018</creationdate><recordtype>article</recordtype><sourceid>DOA</sourceid><recordid>eNqFkc9u3CAQh62qlZqmfYDe_AJshz82prd2laQrRcol7RURGDasvBABTrVvX5ytcmwPiGHE9400v677TGFDQagvASuxz2HDgE6bgQJ_011QOVCiRgFvX2vO3ncfSjkADKNS4qKrV_HRRIuuPy5zDcSZagrWvmYTi8fcz2hyDHHfH7E-Jtf7lPslluUJ83MojWtFSbHPSILDWIMP1tTQOktZMZtIc4UXRWlVxf3pY_fOm7ngp7_3Zffz-up--4Pc3t3stt9uieUcJLEemKFOTpxTMAOjAiYU3hupuBhZe_CRjvDwYMbBo3MSrBwdlW0hUgkL_LLbnb0umYN-yuFo8kknE_RLI-W9NrkGO6N2oMAzisqCF8oO0-SYkQMbnADbTnPRs8vmVEpG_-qjoNcIdItAtwj0GoFeI2jM1zPzO8x4-j-gt7927Pt1C4fKBpMzvH47pCXHtqp_DPsDXtGeHw</recordid><startdate>201812</startdate><enddate>201812</enddate><creator>Xian, Yuqiao</creator><creator>Hu, Haifeng</creator><general>The Institution of Engineering and Technology</general><general>Wiley</general><scope>AAYXX</scope><scope>CITATION</scope><scope>DOA</scope></search><sort><creationdate>201812</creationdate><title>Enhanced multi-dataset transfer learning method for unsupervised person re-identification using co-training strategy</title><author>Xian, Yuqiao ; Hu, Haifeng</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c3307-cf02a1d783310a521408e4ffa79346208e36160bba65fedd70c76d17104794c03</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2018</creationdate><topic>CNN models</topic><topic>co-training strategy</topic><topic>discriminative person representations</topic><topic>enhanced multidataset transfer learning method</topic><topic>image recognition</topic><topic>image representation</topic><topic>iterative methods</topic><topic>iterative training process</topic><topic>labelled source datasets</topic><topic>large-scale benchmark datasets</topic><topic>multiple convolutional neural network models</topic><topic>multiple source datasets</topic><topic>neural nets</topic><topic>pattern clustering</topic><topic>progressive unsupervised co-learning</topic><topic>Research Article</topic><topic>single model</topic><topic>soft labels</topic><topic>target dataset clustering</topic><topic>transferred models</topic><topic>unsupervised learning</topic><topic>unsupervised person re-identification</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Xian, Yuqiao</creatorcontrib><creatorcontrib>Hu, Haifeng</creatorcontrib><collection>CrossRef</collection><collection>Directory of Open Access Journals</collection><jtitle>IET computer vision</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Xian, Yuqiao</au><au>Hu, Haifeng</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Enhanced multi-dataset transfer learning method for unsupervised person re-identification using co-training strategy</atitle><jtitle>IET computer vision</jtitle><date>2018-12</date><risdate>2018</risdate><volume>12</volume><issue>8</issue><spage>1219</spage><epage>1227</epage><pages>1219-1227</pages><issn>1751-9632</issn><issn>1751-9640</issn><eissn>1751-9640</eissn><abstract>This study proposes progressive unsupervised co-learning for unsupervised person re-identification by introducing a co-training strategy in an iterative training process. The authors’ method adopts an iterative training process to improve transferred models by iterating among clustering, selection, exchange, and fine-tuning. To solve the problem of transferring representations learned from multiple source datasets, their method utilises multiple convolutional neural network (CNN) models trained on different labelled source datasets by feeding soft labels obtained by clustering on target dataset to each other. The enhanced model can learn more discriminative person representations than the single model trained on multiple datasets. Experimental results on two large-scale benchmark datasets (i.e. DukeMTMC-reID and Market-1501) demonstrate that their method can enhance transferred CNN models by using more source datasets and is competitive to the state-of-the-art methods.</abstract><pub>The Institution of Engineering and Technology</pub><doi>10.1049/iet-cvi.2018.5103</doi><tpages>9</tpages><oa>free_for_read</oa></addata></record> |
fulltext | fulltext_linktorsrc |
identifier | ISSN: 1751-9632 |
ispartof | IET computer vision, 2018-12, Vol.12 (8), p.1219-1227 |
issn | 1751-9632 1751-9640 1751-9640 |
language | eng |
recordid | cdi_doaj_primary_oai_doaj_org_article_d090f21e9c0f49c588d2a7525d40cd40 |
source | Wiley Online Library Open Access |
subjects | CNN models co-training strategy discriminative person representations enhanced multidataset transfer learning method image recognition image representation iterative methods iterative training process labelled source datasets large-scale benchmark datasets multiple convolutional neural network models multiple source datasets neural nets pattern clustering progressive unsupervised co-learning Research Article single model soft labels target dataset clustering transferred models unsupervised learning unsupervised person re-identification |
title | Enhanced multi-dataset transfer learning method for unsupervised person re-identification using co-training strategy |
url | http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-23T07%3A37%3A23IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-wiley_24P&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Enhanced%20multi-dataset%20transfer%20learning%20method%20for%20unsupervised%20person%20re-identification%20using%20co-training%20strategy&rft.jtitle=IET%20computer%20vision&rft.au=Xian,%20Yuqiao&rft.date=2018-12&rft.volume=12&rft.issue=8&rft.spage=1219&rft.epage=1227&rft.pages=1219-1227&rft.issn=1751-9632&rft.eissn=1751-9640&rft_id=info:doi/10.1049/iet-cvi.2018.5103&rft_dat=%3Cwiley_24P%3ECVI2BF00517%3C/wiley_24P%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c3307-cf02a1d783310a521408e4ffa79346208e36160bba65fedd70c76d17104794c03%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |