No Cover Image

Journal article 206 views 72 downloads

Learning Resolution-Adaptive Representations for Cross-Resolution Person Re-Identification

Yuanbo Wu, Lingqiao Liu Orcid Logo, Yang Wang Orcid Logo, Zheng Zhang Orcid Logo, Farid Boussaid, Mohammed Bennamoun Orcid Logo, Xianghua Xie Orcid Logo

IEEE Transactions on Image Processing, Volume: 32, Pages: 4800 - 4811

Swansea University Authors: Yuanbo Wu, Xianghua Xie Orcid Logo

Check full text

DOI (Published version): 10.1109/tip.2023.3305817

Abstract

Cross-resolution person re-identification (CRReID) is a challenging and practical problem that involves matching low-resolution (LR) query identity images against high-resolution (HR) gallery images. Query images often suffer from resolution degradation due to the different capturing conditions from...

Full description

Published in: IEEE Transactions on Image Processing
ISSN: 1057-7149 1941-0042
Published: Institute of Electrical and Electronics Engineers (IEEE) 2023
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa64124
Tags: Add Tag
No Tags, Be the first to tag this record!
first_indexed 2023-08-24T09:21:08Z
last_indexed 2023-08-24T09:21:08Z
id cronfa64124
recordtype SURis
fullrecord <?xml version="1.0" encoding="utf-8"?><rfc1807 xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema"><bib-version>v2</bib-version><id>64124</id><entry>2023-08-24</entry><title>Learning Resolution-Adaptive Representations for Cross-Resolution Person Re-Identification</title><swanseaauthors><author><sid>205b1ac5a767e977bebb5d6afd770784</sid><firstname>Yuanbo</firstname><surname>Wu</surname><name>Yuanbo Wu</name><active>true</active><ethesisStudent>false</ethesisStudent></author><author><sid>b334d40963c7a2f435f06d2c26c74e11</sid><ORCID>0000-0002-2701-8660</ORCID><firstname>Xianghua</firstname><surname>Xie</surname><name>Xianghua Xie</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2023-08-24</date><deptcode>SCS</deptcode><abstract>Cross-resolution person re-identification (CRReID) is a challenging and practical problem that involves matching low-resolution (LR) query identity images against high-resolution (HR) gallery images. Query images often suffer from resolution degradation due to the different capturing conditions from real-world cameras. State-of-the-art solutions for CRReID either learn a resolution-invariant representation or adopt a super-resolution (SR) module to recover the missing information from the LR query. In this paper, we propose an alternative SR-free paradigm to directly compare HR and LR images via a dynamic metric that is adaptive to the resolution of a query image. We realize this idea by learning resolution-adaptive representations for cross-resolution comparison. We propose two resolution-adaptive mechanisms to achieve this. The first mechanism encodes the resolution specifics into different subvectors in the penultimate layer of the deep neural network, creating a varying-length representation. To better extract resolution-dependent information, we further propose to learn resolution-adaptive masks for intermediate residual feature blocks. A novel progressive learning strategy is proposed to train those masks properly. These two mechanisms are combined to boost the performance of CRReID. Experimental results show that the proposed method outperforms existing approaches and achieves state-of-the-art performance.</abstract><type>Journal Article</type><journal>IEEE Transactions on Image Processing</journal><volume>32</volume><journalNumber/><paginationStart>4800</paginationStart><paginationEnd>4811</paginationEnd><publisher>Institute of Electrical and Electronics Engineers (IEEE)</publisher><placeOfPublication/><isbnPrint/><isbnElectronic/><issnPrint>1057-7149</issnPrint><issnElectronic>1941-0042</issnElectronic><keywords>Image resolution, Feature extraction ,Measurement,Training, Superresolution, Cameras, Adaptation models</keywords><publishedDay>23</publishedDay><publishedMonth>8</publishedMonth><publishedYear>2023</publishedYear><publishedDate>2023-08-23</publishedDate><doi>10.1109/tip.2023.3305817</doi><url>http://dx.doi.org/10.1109/tip.2023.3305817</url><notes/><college>COLLEGE NANME</college><department>Computer Science</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>SCS</DepartmentCode><institution>Swansea University</institution><apcterm>Not Required</apcterm><funders/><projectreference/><lastEdited>2023-10-02T11:01:37.9904970</lastEdited><Created>2023-08-24T10:15:57.9692312</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Mathematics and Computer Science - Computer Science</level></path><authors><author><firstname>Yuanbo</firstname><surname>Wu</surname><order>1</order></author><author><firstname>Lingqiao</firstname><surname>Liu</surname><orcid>0000-0003-3584-795x</orcid><order>2</order></author><author><firstname>Yang</firstname><surname>Wang</surname><orcid>0000-0003-1029-9280</orcid><order>3</order></author><author><firstname>Zheng</firstname><surname>Zhang</surname><orcid>0000-0003-1470-6998</orcid><order>4</order></author><author><firstname>Farid</firstname><surname>Boussaid</surname><order>5</order></author><author><firstname>Mohammed</firstname><surname>Bennamoun</surname><orcid>0000-0002-6603-3257</orcid><order>6</order></author><author><firstname>Xianghua</firstname><surname>Xie</surname><orcid>0000-0002-2701-8660</orcid><order>7</order></author></authors><documents><document><filename>64124__28387__cedb4034b25847e4b06f69907c31fd30.pdf</filename><originalFilename>Cross_Resolution_ReID_IEEE_Trans.pdf</originalFilename><uploaded>2023-08-28T10:36:53.9142498</uploaded><type>Output</type><contentLength>779558</contentLength><contentType>application/pdf</contentType><version>Accepted Manuscript</version><cronfaStatus>true</cronfaStatus><copyrightCorrect>true</copyrightCorrect><language>eng</language></document></documents><OutputDurs/></rfc1807>
spelling v2 64124 2023-08-24 Learning Resolution-Adaptive Representations for Cross-Resolution Person Re-Identification 205b1ac5a767e977bebb5d6afd770784 Yuanbo Wu Yuanbo Wu true false b334d40963c7a2f435f06d2c26c74e11 0000-0002-2701-8660 Xianghua Xie Xianghua Xie true false 2023-08-24 SCS Cross-resolution person re-identification (CRReID) is a challenging and practical problem that involves matching low-resolution (LR) query identity images against high-resolution (HR) gallery images. Query images often suffer from resolution degradation due to the different capturing conditions from real-world cameras. State-of-the-art solutions for CRReID either learn a resolution-invariant representation or adopt a super-resolution (SR) module to recover the missing information from the LR query. In this paper, we propose an alternative SR-free paradigm to directly compare HR and LR images via a dynamic metric that is adaptive to the resolution of a query image. We realize this idea by learning resolution-adaptive representations for cross-resolution comparison. We propose two resolution-adaptive mechanisms to achieve this. The first mechanism encodes the resolution specifics into different subvectors in the penultimate layer of the deep neural network, creating a varying-length representation. To better extract resolution-dependent information, we further propose to learn resolution-adaptive masks for intermediate residual feature blocks. A novel progressive learning strategy is proposed to train those masks properly. These two mechanisms are combined to boost the performance of CRReID. Experimental results show that the proposed method outperforms existing approaches and achieves state-of-the-art performance. Journal Article IEEE Transactions on Image Processing 32 4800 4811 Institute of Electrical and Electronics Engineers (IEEE) 1057-7149 1941-0042 Image resolution, Feature extraction ,Measurement,Training, Superresolution, Cameras, Adaptation models 23 8 2023 2023-08-23 10.1109/tip.2023.3305817 http://dx.doi.org/10.1109/tip.2023.3305817 COLLEGE NANME Computer Science COLLEGE CODE SCS Swansea University Not Required 2023-10-02T11:01:37.9904970 2023-08-24T10:15:57.9692312 Faculty of Science and Engineering School of Mathematics and Computer Science - Computer Science Yuanbo Wu 1 Lingqiao Liu 0000-0003-3584-795x 2 Yang Wang 0000-0003-1029-9280 3 Zheng Zhang 0000-0003-1470-6998 4 Farid Boussaid 5 Mohammed Bennamoun 0000-0002-6603-3257 6 Xianghua Xie 0000-0002-2701-8660 7 64124__28387__cedb4034b25847e4b06f69907c31fd30.pdf Cross_Resolution_ReID_IEEE_Trans.pdf 2023-08-28T10:36:53.9142498 Output 779558 application/pdf Accepted Manuscript true true eng
title Learning Resolution-Adaptive Representations for Cross-Resolution Person Re-Identification
spellingShingle Learning Resolution-Adaptive Representations for Cross-Resolution Person Re-Identification
Yuanbo Wu
Xianghua Xie
title_short Learning Resolution-Adaptive Representations for Cross-Resolution Person Re-Identification
title_full Learning Resolution-Adaptive Representations for Cross-Resolution Person Re-Identification
title_fullStr Learning Resolution-Adaptive Representations for Cross-Resolution Person Re-Identification
title_full_unstemmed Learning Resolution-Adaptive Representations for Cross-Resolution Person Re-Identification
title_sort Learning Resolution-Adaptive Representations for Cross-Resolution Person Re-Identification
author_id_str_mv 205b1ac5a767e977bebb5d6afd770784
b334d40963c7a2f435f06d2c26c74e11
author_id_fullname_str_mv 205b1ac5a767e977bebb5d6afd770784_***_Yuanbo Wu
b334d40963c7a2f435f06d2c26c74e11_***_Xianghua Xie
author Yuanbo Wu
Xianghua Xie
author2 Yuanbo Wu
Lingqiao Liu
Yang Wang
Zheng Zhang
Farid Boussaid
Mohammed Bennamoun
Xianghua Xie
format Journal article
container_title IEEE Transactions on Image Processing
container_volume 32
container_start_page 4800
publishDate 2023
institution Swansea University
issn 1057-7149
1941-0042
doi_str_mv 10.1109/tip.2023.3305817
publisher Institute of Electrical and Electronics Engineers (IEEE)
college_str Faculty of Science and Engineering
hierarchytype
hierarchy_top_id facultyofscienceandengineering
hierarchy_top_title Faculty of Science and Engineering
hierarchy_parent_id facultyofscienceandengineering
hierarchy_parent_title Faculty of Science and Engineering
department_str School of Mathematics and Computer Science - Computer Science{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Mathematics and Computer Science - Computer Science
url http://dx.doi.org/10.1109/tip.2023.3305817
document_store_str 1
active_str 0
description Cross-resolution person re-identification (CRReID) is a challenging and practical problem that involves matching low-resolution (LR) query identity images against high-resolution (HR) gallery images. Query images often suffer from resolution degradation due to the different capturing conditions from real-world cameras. State-of-the-art solutions for CRReID either learn a resolution-invariant representation or adopt a super-resolution (SR) module to recover the missing information from the LR query. In this paper, we propose an alternative SR-free paradigm to directly compare HR and LR images via a dynamic metric that is adaptive to the resolution of a query image. We realize this idea by learning resolution-adaptive representations for cross-resolution comparison. We propose two resolution-adaptive mechanisms to achieve this. The first mechanism encodes the resolution specifics into different subvectors in the penultimate layer of the deep neural network, creating a varying-length representation. To better extract resolution-dependent information, we further propose to learn resolution-adaptive masks for intermediate residual feature blocks. A novel progressive learning strategy is proposed to train those masks properly. These two mechanisms are combined to boost the performance of CRReID. Experimental results show that the proposed method outperforms existing approaches and achieves state-of-the-art performance.
published_date 2023-08-23T11:01:41Z
_version_ 1778637499128610816
score 11.01306

Similar Items