No Cover Image

Journal article 387 views 399 downloads

Class activation map guided level sets for weakly supervised semantic segmentation

Yifan Wang, Gerald Schaefer, Xiyao Liu, Jing Dong Orcid Logo, Linglin Jing, Ye Wei Orcid Logo, Xianghua Xie Orcid Logo, Hui Fang Orcid Logo

Pattern Recognition, Volume: 165, Start page: 111566

Swansea University Author: Xianghua Xie Orcid Logo

  • 69059.pdf

    PDF | Accepted Manuscript

    Author accepted manuscript document released under the terms of a Creative Commons CC-BY licence using the Swansea University Research Publications Policy (rights retention).

    Download (3.13MB)

Check full text

DOI (Published version): 10.1016/j.patcog.2025.111566

Abstract

Weakly supervised semantic segmentation (WSSS) aims to achieve pixel-level fine-grained image segmentation using only weak guidance such as image level class labels, thus significantly decreasing annotation costs. Despite the impressive performance showcased by current state-of-the-art WSSS approach...

Full description

Published in: Pattern Recognition
ISSN: 0031-3203
Published: Elsevier BV 2025
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa69059
first_indexed 2025-03-07T08:43:46Z
last_indexed 2025-04-12T04:36:10Z
id cronfa69059
recordtype SURis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2025-04-11T12:22:46.2265527</datestamp><bib-version>v2</bib-version><id>69059</id><entry>2025-03-07</entry><title>Class activation map guided level sets for weakly supervised semantic segmentation</title><swanseaauthors><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>2025-03-07</date><deptcode>MACS</deptcode><abstract>Weakly supervised semantic segmentation (WSSS) aims to achieve pixel-level fine-grained image segmentation using only weak guidance such as image level class labels, thus significantly decreasing annotation costs. Despite the impressive performance showcased by current state-of-the-art WSSS approaches, the lack of precise object localisation limits their segmentation accuracy, especially for pixels close to object boundaries. To address this issue, we propose a novel class activation map (CAM)-based level set method to effectively improve the quality of pseudo-labels by exploring the capabilityof level sets to enhance the segmentation accuracy at object boundaries. To speed up the level set evolution process, we use Fourier neural operators to simulate the dynamic evolution of our level set method. Extensive experimental results show that our approach significantly outperforms existingWSSS methods on both PASCAL VOC 2012 and MS COCO datasets.</abstract><type>Journal Article</type><journal>Pattern Recognition</journal><volume>165</volume><journalNumber/><paginationStart>111566</paginationStart><paginationEnd/><publisher>Elsevier BV</publisher><placeOfPublication/><isbnPrint/><isbnElectronic/><issnPrint>0031-3203</issnPrint><issnElectronic/><keywords>Weakly supervised semantic segmentation; Class activation map; Pseudo-label; Level set; Fourier neural operator</keywords><publishedDay>1</publishedDay><publishedMonth>9</publishedMonth><publishedYear>2025</publishedYear><publishedDate>2025-09-01</publishedDate><doi>10.1016/j.patcog.2025.111566</doi><url/><notes/><college>COLLEGE NANME</college><department>Mathematics and Computer Science School</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>MACS</DepartmentCode><institution>Swansea University</institution><apcterm/><funders>This research is supported by Natural Science Foundation of Hunan Province, China (2022GK5002, 2024JK2015, 2024JJ5440), 111 Project (D23006), Special Foundation for Distinguished Young Scientists of Changsha (kq2209003), Dalian Major Projects of Basic Research (2023JJ11CG002) and Interdisciplinary Research Project of Dalian University (DLUXK-2024-YB-007).</funders><projectreference/><lastEdited>2025-04-11T12:22:46.2265527</lastEdited><Created>2025-03-07T08:39:55.3925788</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>Yifan</firstname><surname>Wang</surname><order>1</order></author><author><firstname>Gerald</firstname><surname>Schaefer</surname><order>2</order></author><author><firstname>Xiyao</firstname><surname>Liu</surname><order>3</order></author><author><firstname>Jing</firstname><surname>Dong</surname><orcid>0000-0003-3489-6661</orcid><order>4</order></author><author><firstname>Linglin</firstname><surname>Jing</surname><order>5</order></author><author><firstname>Ye</firstname><surname>Wei</surname><orcid>0009-0003-1842-5867</orcid><order>6</order></author><author><firstname>Xianghua</firstname><surname>Xie</surname><orcid>0000-0002-2701-8660</orcid><order>7</order></author><author><firstname>Hui</firstname><surname>Fang</surname><orcid>0000-0001-9365-7420</orcid><order>8</order></author></authors><documents><document><filename>69059__33757__1d264bf9c8034c3bb7d30dccd173468e.pdf</filename><originalFilename>69059.pdf</originalFilename><uploaded>2025-03-07T08:43:39.0008252</uploaded><type>Output</type><contentLength>3281440</contentLength><contentType>application/pdf</contentType><version>Accepted Manuscript</version><cronfaStatus>true</cronfaStatus><documentNotes>Author accepted manuscript document released under the terms of a Creative Commons CC-BY licence using the Swansea University Research Publications Policy (rights retention).</documentNotes><copyrightCorrect>true</copyrightCorrect><language>eng</language><licence>https://creativecommons.org/licenses/by/4.0/deed.en</licence></document></documents><OutputDurs/></rfc1807>
spelling 2025-04-11T12:22:46.2265527 v2 69059 2025-03-07 Class activation map guided level sets for weakly supervised semantic segmentation b334d40963c7a2f435f06d2c26c74e11 0000-0002-2701-8660 Xianghua Xie Xianghua Xie true false 2025-03-07 MACS Weakly supervised semantic segmentation (WSSS) aims to achieve pixel-level fine-grained image segmentation using only weak guidance such as image level class labels, thus significantly decreasing annotation costs. Despite the impressive performance showcased by current state-of-the-art WSSS approaches, the lack of precise object localisation limits their segmentation accuracy, especially for pixels close to object boundaries. To address this issue, we propose a novel class activation map (CAM)-based level set method to effectively improve the quality of pseudo-labels by exploring the capabilityof level sets to enhance the segmentation accuracy at object boundaries. To speed up the level set evolution process, we use Fourier neural operators to simulate the dynamic evolution of our level set method. Extensive experimental results show that our approach significantly outperforms existingWSSS methods on both PASCAL VOC 2012 and MS COCO datasets. Journal Article Pattern Recognition 165 111566 Elsevier BV 0031-3203 Weakly supervised semantic segmentation; Class activation map; Pseudo-label; Level set; Fourier neural operator 1 9 2025 2025-09-01 10.1016/j.patcog.2025.111566 COLLEGE NANME Mathematics and Computer Science School COLLEGE CODE MACS Swansea University This research is supported by Natural Science Foundation of Hunan Province, China (2022GK5002, 2024JK2015, 2024JJ5440), 111 Project (D23006), Special Foundation for Distinguished Young Scientists of Changsha (kq2209003), Dalian Major Projects of Basic Research (2023JJ11CG002) and Interdisciplinary Research Project of Dalian University (DLUXK-2024-YB-007). 2025-04-11T12:22:46.2265527 2025-03-07T08:39:55.3925788 Faculty of Science and Engineering School of Mathematics and Computer Science - Computer Science Yifan Wang 1 Gerald Schaefer 2 Xiyao Liu 3 Jing Dong 0000-0003-3489-6661 4 Linglin Jing 5 Ye Wei 0009-0003-1842-5867 6 Xianghua Xie 0000-0002-2701-8660 7 Hui Fang 0000-0001-9365-7420 8 69059__33757__1d264bf9c8034c3bb7d30dccd173468e.pdf 69059.pdf 2025-03-07T08:43:39.0008252 Output 3281440 application/pdf Accepted Manuscript true Author accepted manuscript document released under the terms of a Creative Commons CC-BY licence using the Swansea University Research Publications Policy (rights retention). true eng https://creativecommons.org/licenses/by/4.0/deed.en
title Class activation map guided level sets for weakly supervised semantic segmentation
spellingShingle Class activation map guided level sets for weakly supervised semantic segmentation
Xianghua Xie
title_short Class activation map guided level sets for weakly supervised semantic segmentation
title_full Class activation map guided level sets for weakly supervised semantic segmentation
title_fullStr Class activation map guided level sets for weakly supervised semantic segmentation
title_full_unstemmed Class activation map guided level sets for weakly supervised semantic segmentation
title_sort Class activation map guided level sets for weakly supervised semantic segmentation
author_id_str_mv b334d40963c7a2f435f06d2c26c74e11
author_id_fullname_str_mv b334d40963c7a2f435f06d2c26c74e11_***_Xianghua Xie
author Xianghua Xie
author2 Yifan Wang
Gerald Schaefer
Xiyao Liu
Jing Dong
Linglin Jing
Ye Wei
Xianghua Xie
Hui Fang
format Journal article
container_title Pattern Recognition
container_volume 165
container_start_page 111566
publishDate 2025
institution Swansea University
issn 0031-3203
doi_str_mv 10.1016/j.patcog.2025.111566
publisher Elsevier BV
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
document_store_str 1
active_str 0
description Weakly supervised semantic segmentation (WSSS) aims to achieve pixel-level fine-grained image segmentation using only weak guidance such as image level class labels, thus significantly decreasing annotation costs. Despite the impressive performance showcased by current state-of-the-art WSSS approaches, the lack of precise object localisation limits their segmentation accuracy, especially for pixels close to object boundaries. To address this issue, we propose a novel class activation map (CAM)-based level set method to effectively improve the quality of pseudo-labels by exploring the capabilityof level sets to enhance the segmentation accuracy at object boundaries. To speed up the level set evolution process, we use Fourier neural operators to simulate the dynamic evolution of our level set method. Extensive experimental results show that our approach significantly outperforms existingWSSS methods on both PASCAL VOC 2012 and MS COCO datasets.
published_date 2025-09-01T05:27:11Z
_version_ 1851097802560503808
score 11.089407

Similar Items