Self-Regulation Feature Network for Person Reidentification

The performance of person reidentification (ReID) has been significantly improved with the development of convolutional neural networks in the past few years. Most existing ReID methods focus on modifying the backbone network architecture to learn feature representations of persons; however, there a...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on instrumentation and measurement 2023, Vol.72, p.1-8
Hauptverfasser: Tian, Hui, Hu, Junlin
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page 8
container_issue
container_start_page 1
container_title IEEE transactions on instrumentation and measurement
container_volume 72
creator Tian, Hui
Hu, Junlin
description The performance of person reidentification (ReID) has been significantly improved with the development of convolutional neural networks in the past few years. Most existing ReID methods focus on modifying the backbone network architecture to learn feature representations of persons; however, there are only a few methods that explore the potential performance of the network architecture to relearn features. In this article, we propose a self-regulation feature (SRF) scheme based on the network model, called SRFnet, to address the problem of the limited ability of the model for learning person features. The proposed SRFnet utilizes global branches to supervise local branches and deeply mine potential features of the network from the perspective of optimization to obtain more distinguished feature representations of persons. Different from methods that add an attention mechanism, our SRFnet can self-adjust feature learning by only adding loss functions between the local and global features of the network, and it does not require any additional convolution parameters. Extensive experiments on three public datasets show that the proposed SRFnet method achieves the results of mean average precision (mAP)/Rank-1 as 93.65%/95.28% on Market1501, 89.65%/91.79% on DuckMTMC-reID, 83.79%/82.00% on CUHK03 (labeled), and 81.92%/80.21% on CUHK03 (detected) datasets, respectively. Furthermore, experimental results on cross-domain datasets also demonstrate the effectiveness of the proposed SRFnet.
doi_str_mv 10.1109/TIM.2023.3240217
format Article
fullrecord <record><control><sourceid>proquest_RIE</sourceid><recordid>TN_cdi_crossref_primary_10_1109_TIM_2023_3240217</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>10026824</ieee_id><sourcerecordid>2774319123</sourcerecordid><originalsourceid>FETCH-LOGICAL-c292t-ce23a3396288f8c69da77abfee425eca190200f79d9c69ccb159f3f04650a2b33</originalsourceid><addsrcrecordid>eNpNkM9PAjEQhRujiYjePXjYxPPidNptt_FkiCgJ_gjiuSllahaRxXY3hv_eRTh4msP73pvkY-ySw4BzMDez8dMAAcVAoATk-oj1eFHo3CiFx6wHwMvcyEKdsrOUlgCgldQ9dvtGq5BP6aNduaaq19mIXNNGyp6p-anjZxbqmL1STF00pWpB66YKlf9jz9lJcKtEF4fbZ--j-9nwMZ-8PIyHd5Pco8Em94TCCWEUlmUovTILp7WbByKJBXnHDSBA0GZhutD7OS9MEAGkKsDhXIg-u97vbmL93VJq7LJu47p7aVFrKbjhuKNgT_lYpxQp2E2svlzcWg52p8h2iuxOkT0o6ipX-0pFRP9wQFWiFL8ly2Gl</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2774319123</pqid></control><display><type>article</type><title>Self-Regulation Feature Network for Person Reidentification</title><source>IEEE Electronic Library (IEL)</source><creator>Tian, Hui ; Hu, Junlin</creator><creatorcontrib>Tian, Hui ; Hu, Junlin</creatorcontrib><description>The performance of person reidentification (ReID) has been significantly improved with the development of convolutional neural networks in the past few years. Most existing ReID methods focus on modifying the backbone network architecture to learn feature representations of persons; however, there are only a few methods that explore the potential performance of the network architecture to relearn features. In this article, we propose a self-regulation feature (SRF) scheme based on the network model, called SRFnet, to address the problem of the limited ability of the model for learning person features. The proposed SRFnet utilizes global branches to supervise local branches and deeply mine potential features of the network from the perspective of optimization to obtain more distinguished feature representations of persons. Different from methods that add an attention mechanism, our SRFnet can self-adjust feature learning by only adding loss functions between the local and global features of the network, and it does not require any additional convolution parameters. Extensive experiments on three public datasets show that the proposed SRFnet method achieves the results of mean average precision (mAP)/Rank-1 as 93.65%/95.28% on Market1501, 89.65%/91.79% on DuckMTMC-reID, 83.79%/82.00% on CUHK03 (labeled), and 81.92%/80.21% on CUHK03 (detected) datasets, respectively. Furthermore, experimental results on cross-domain datasets also demonstrate the effectiveness of the proposed SRFnet.</description><identifier>ISSN: 0018-9456</identifier><identifier>EISSN: 1557-9662</identifier><identifier>DOI: 10.1109/TIM.2023.3240217</identifier><identifier>CODEN: IEIMAO</identifier><language>eng</language><publisher>New York: IEEE</publisher><subject>Artificial neural networks ; Computer architecture ; Computer networks ; Convolution ; Convolutional neural networks ; Datasets ; Deep learning ; Feature extraction ; feature learning ; feature representation ; Neural networks ; Optimization ; person reidentification (ReID) ; Representations ; self-regulation ; Task analysis ; Training</subject><ispartof>IEEE transactions on instrumentation and measurement, 2023, Vol.72, p.1-8</ispartof><rights>Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2023</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c292t-ce23a3396288f8c69da77abfee425eca190200f79d9c69ccb159f3f04650a2b33</citedby><cites>FETCH-LOGICAL-c292t-ce23a3396288f8c69da77abfee425eca190200f79d9c69ccb159f3f04650a2b33</cites><orcidid>0000-0001-7394-8466 ; 0000-0002-0117-3494</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/10026824$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>314,780,784,796,4024,27923,27924,27925,54758</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/10026824$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Tian, Hui</creatorcontrib><creatorcontrib>Hu, Junlin</creatorcontrib><title>Self-Regulation Feature Network for Person Reidentification</title><title>IEEE transactions on instrumentation and measurement</title><addtitle>TIM</addtitle><description>The performance of person reidentification (ReID) has been significantly improved with the development of convolutional neural networks in the past few years. Most existing ReID methods focus on modifying the backbone network architecture to learn feature representations of persons; however, there are only a few methods that explore the potential performance of the network architecture to relearn features. In this article, we propose a self-regulation feature (SRF) scheme based on the network model, called SRFnet, to address the problem of the limited ability of the model for learning person features. The proposed SRFnet utilizes global branches to supervise local branches and deeply mine potential features of the network from the perspective of optimization to obtain more distinguished feature representations of persons. Different from methods that add an attention mechanism, our SRFnet can self-adjust feature learning by only adding loss functions between the local and global features of the network, and it does not require any additional convolution parameters. Extensive experiments on three public datasets show that the proposed SRFnet method achieves the results of mean average precision (mAP)/Rank-1 as 93.65%/95.28% on Market1501, 89.65%/91.79% on DuckMTMC-reID, 83.79%/82.00% on CUHK03 (labeled), and 81.92%/80.21% on CUHK03 (detected) datasets, respectively. Furthermore, experimental results on cross-domain datasets also demonstrate the effectiveness of the proposed SRFnet.</description><subject>Artificial neural networks</subject><subject>Computer architecture</subject><subject>Computer networks</subject><subject>Convolution</subject><subject>Convolutional neural networks</subject><subject>Datasets</subject><subject>Deep learning</subject><subject>Feature extraction</subject><subject>feature learning</subject><subject>feature representation</subject><subject>Neural networks</subject><subject>Optimization</subject><subject>person reidentification (ReID)</subject><subject>Representations</subject><subject>self-regulation</subject><subject>Task analysis</subject><subject>Training</subject><issn>0018-9456</issn><issn>1557-9662</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>RIE</sourceid><recordid>eNpNkM9PAjEQhRujiYjePXjYxPPidNptt_FkiCgJ_gjiuSllahaRxXY3hv_eRTh4msP73pvkY-ySw4BzMDez8dMAAcVAoATk-oj1eFHo3CiFx6wHwMvcyEKdsrOUlgCgldQ9dvtGq5BP6aNduaaq19mIXNNGyp6p-anjZxbqmL1STF00pWpB66YKlf9jz9lJcKtEF4fbZ--j-9nwMZ-8PIyHd5Pco8Em94TCCWEUlmUovTILp7WbByKJBXnHDSBA0GZhutD7OS9MEAGkKsDhXIg-u97vbmL93VJq7LJu47p7aVFrKbjhuKNgT_lYpxQp2E2svlzcWg52p8h2iuxOkT0o6ipX-0pFRP9wQFWiFL8ly2Gl</recordid><startdate>2023</startdate><enddate>2023</enddate><creator>Tian, Hui</creator><creator>Hu, Junlin</creator><general>IEEE</general><general>The Institute of Electrical and Electronics Engineers, Inc. (IEEE)</general><scope>97E</scope><scope>RIA</scope><scope>RIE</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7SP</scope><scope>7U5</scope><scope>8FD</scope><scope>L7M</scope><orcidid>https://orcid.org/0000-0001-7394-8466</orcidid><orcidid>https://orcid.org/0000-0002-0117-3494</orcidid></search><sort><creationdate>2023</creationdate><title>Self-Regulation Feature Network for Person Reidentification</title><author>Tian, Hui ; Hu, Junlin</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c292t-ce23a3396288f8c69da77abfee425eca190200f79d9c69ccb159f3f04650a2b33</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Artificial neural networks</topic><topic>Computer architecture</topic><topic>Computer networks</topic><topic>Convolution</topic><topic>Convolutional neural networks</topic><topic>Datasets</topic><topic>Deep learning</topic><topic>Feature extraction</topic><topic>feature learning</topic><topic>feature representation</topic><topic>Neural networks</topic><topic>Optimization</topic><topic>person reidentification (ReID)</topic><topic>Representations</topic><topic>self-regulation</topic><topic>Task analysis</topic><topic>Training</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Tian, Hui</creatorcontrib><creatorcontrib>Hu, Junlin</creatorcontrib><collection>IEEE All-Society Periodicals Package (ASPP) 2005-present</collection><collection>IEEE All-Society Periodicals Package (ASPP) 1998-Present</collection><collection>IEEE Electronic Library (IEL)</collection><collection>CrossRef</collection><collection>Electronics &amp; Communications Abstracts</collection><collection>Solid State and Superconductivity Abstracts</collection><collection>Technology Research Database</collection><collection>Advanced Technologies Database with Aerospace</collection><jtitle>IEEE transactions on instrumentation and measurement</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Tian, Hui</au><au>Hu, Junlin</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Self-Regulation Feature Network for Person Reidentification</atitle><jtitle>IEEE transactions on instrumentation and measurement</jtitle><stitle>TIM</stitle><date>2023</date><risdate>2023</risdate><volume>72</volume><spage>1</spage><epage>8</epage><pages>1-8</pages><issn>0018-9456</issn><eissn>1557-9662</eissn><coden>IEIMAO</coden><abstract>The performance of person reidentification (ReID) has been significantly improved with the development of convolutional neural networks in the past few years. Most existing ReID methods focus on modifying the backbone network architecture to learn feature representations of persons; however, there are only a few methods that explore the potential performance of the network architecture to relearn features. In this article, we propose a self-regulation feature (SRF) scheme based on the network model, called SRFnet, to address the problem of the limited ability of the model for learning person features. The proposed SRFnet utilizes global branches to supervise local branches and deeply mine potential features of the network from the perspective of optimization to obtain more distinguished feature representations of persons. Different from methods that add an attention mechanism, our SRFnet can self-adjust feature learning by only adding loss functions between the local and global features of the network, and it does not require any additional convolution parameters. Extensive experiments on three public datasets show that the proposed SRFnet method achieves the results of mean average precision (mAP)/Rank-1 as 93.65%/95.28% on Market1501, 89.65%/91.79% on DuckMTMC-reID, 83.79%/82.00% on CUHK03 (labeled), and 81.92%/80.21% on CUHK03 (detected) datasets, respectively. Furthermore, experimental results on cross-domain datasets also demonstrate the effectiveness of the proposed SRFnet.</abstract><cop>New York</cop><pub>IEEE</pub><doi>10.1109/TIM.2023.3240217</doi><tpages>8</tpages><orcidid>https://orcid.org/0000-0001-7394-8466</orcidid><orcidid>https://orcid.org/0000-0002-0117-3494</orcidid></addata></record>
fulltext fulltext_linktorsrc
identifier ISSN: 0018-9456
ispartof IEEE transactions on instrumentation and measurement, 2023, Vol.72, p.1-8
issn 0018-9456
1557-9662
language eng
recordid cdi_crossref_primary_10_1109_TIM_2023_3240217
source IEEE Electronic Library (IEL)
subjects Artificial neural networks
Computer architecture
Computer networks
Convolution
Convolutional neural networks
Datasets
Deep learning
Feature extraction
feature learning
feature representation
Neural networks
Optimization
person reidentification (ReID)
Representations
self-regulation
Task analysis
Training
title Self-Regulation Feature Network for Person Reidentification
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-06T16%3A22%3A27IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_RIE&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Self-Regulation%20Feature%20Network%20for%20Person%20Reidentification&rft.jtitle=IEEE%20transactions%20on%20instrumentation%20and%20measurement&rft.au=Tian,%20Hui&rft.date=2023&rft.volume=72&rft.spage=1&rft.epage=8&rft.pages=1-8&rft.issn=0018-9456&rft.eissn=1557-9662&rft.coden=IEIMAO&rft_id=info:doi/10.1109/TIM.2023.3240217&rft_dat=%3Cproquest_RIE%3E2774319123%3C/proquest_RIE%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2774319123&rft_id=info:pmid/&rft_ieee_id=10026824&rfr_iscdi=true