CINXE.COM

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> <html xmlns="http://www.w3.org/1999/xhtml"> <head> <meta http-equiv="Content-Type" content="text/html; charset=utf-8" /> <title>ComSIS | Computer Science and Information Systems</title> <link rel="stylesheet" type="text/css" href="res/style1.css" /> </head> <body> <script type="text/javascript" src="res/wz_tooltip.js"></script> <script type="text/javascript" src="res/slide.js"></script> <div id="all"> <div id="header"> <h1>Computer Science and Information Systems</h1> </div>  <div id="main"> <div id="sidebar"> <p>About the journal</p> <ul> <li><a href="index.php">Home page</a></li> <li><a href="contact.php">Contact information</a></li> <li><a href="aims.php">Aims and scope</a></li> <li><a href="indexing.php">Indexing information</a></li> <li><a href="policies.php">Editorial policies</a></li> <li><a href="consortium.php">ComSIS consortium</a></li> <li><a href="boards.php">Journal boards</a></li> <li><a href="managing.php">Managing board</a></li> </ul> <p>For authors</p> <ul> <li><a href="information.php">Information for contributors</a></li> <li><a href="http://ojs.pmf.uns.ac.rs/index.php/comsis">Paper submission</a></li> <li><a href="submission.php">Article submission through OJS</a></li> <li><a href="copyright.php">Copyright transfer form</a></li> <li><a href="download.php">Download section</a></li> </ul> <p>For readers</p> <ul> <li><a href="archive.php?show=lstnew">Forthcoming articles</a></li> <li><a href="archive.php?show=vol2104">Current issue</a></li> <li><a href="archive.php">Archive</a></li> </ul> <p>For reviewers</p> <ul> <li><a href="http://ojs.pmf.uns.ac.rs/index.php/comsis">View and review submissions</a></li> </ul> <p>News</p> <ul> <li><a href="https://www.facebook.com/ComSISJournal/"> <img src="res/fb.png" alt="FB"/> Journal's Facebook page</a></li> <li><a href="cfp.php">Calls for special issues</a></li> <li><a href="notification.php">New issue notification</a></li> </ul> </div>  <div id="content">  <h1 class="title">Learning Discriminative Representations through an Attention Mechanism for Image-based Person Re-identification</h1><p class="authors">Jing Liu<sup>1, 2</sup> and Guoqing Zhou<sup>2</sup></p><ol><li>School of Computer Science, Weinan Normal University,<br/>Weinan, 714099, Shaanxi, China.<br/>liujing8318@mail.nwpu.edu.cn</li><li>School of Computer Science, Northwestern Polytechnical<br/>University, Xi’an, 710072, Shaanxi, China.<br/>Zhouguoqing@nwpu.edu.cn</li></ol><h3>Abstract</h3><p>Over the past years, person re-identification has been obtaining various attentions in computer vision tasks. However, existing methods mainly focus on building massive number of deep architecture layers, which is unsuitable for extracting the robust features for person re-ID. In this paper, we present a novel hybrid framework PGAN, through which the discriminative representations can be learned for person re-ID. Specifically, a novel self-attention method named channel-wise attention mechanism is adopted to learn the informative representations from the patch-network and global network, respectively. In addition, CSwin Transformer is exploited to re-extract the discriminative features from the residual blocks. We obtain a mAP of 81.8% and 80.3% of the labeled and detected dataset on the CUHK0-NP dataset. And we obtain a mAP of 83.4% and 91.3% on the DukeMTMC and Market-1501 datasets respectively. Comprehensive experiments are performed on the three datasets, (Market-1501, DukeMTMC-reID and CUHK03-NP), demonstrating the efficiency of the introduced approach.</p><h3>Key words</h3><p>person re-identification, channel-wise attention, deep learning</p><h3>Digital Object Identifier (DOI)</h3><p><a href="https://doi.org/10.2298/CSIS230829044L">https://doi.org/10.2298/CSIS230829044L</a></p><h3>Publication information</h3><p><a href="/archive.php?show=vol2104">Volume 21, Issue 4 (September 2024)</a><br/>Year of Publication: 2024<br/>ISSN: 2406-1018 (Online)<br/>Publisher: ComSIS Consortium</p><h3>Full text</h3><p><a class="download" href="pdf.php?id=16451"><img class="left" src="res/pdf.png" alt="Download"/>Available in PDF<br/><em>Portable Document Format</em></a></p><h3>How to cite</h3><p>Liu, J., Zhou, G.: Learning Discriminative Representations through an Attention Mechanism for Image-based Person Re-identification. Computer Science and Information Systems, Vol. 21, No. 4, 1483–1498. (2024), https://doi.org/10.2298/CSIS230829044L</p>  </div>  </div>  <div id="footer_top"> </div> <div id="footer"> <div class="left">Faculty of Sciences, Trg Dositeja Obradovića 3, 21000 Novi Sad, Serbia, <a href="mailto:comsis@uns.ac.rs">comsis@uns.ac.rs</a></div> <div class="left">Published by ComSIS Consortium under<br/><a rel="license" href="http://creativecommons.org/licenses/by-nc-nd/4.0/">Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License<br><img alt="Creative Commons License" style="border-width:0" src="https://i.creativecommons.org/l/by-nc-nd/4.0/88x31.png"/></a></div> <div class="clearer"> </div> </div>  </div>  </body> </html>