No Cover Image

Journal article 66 views

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
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.
Keywords: Weakly supervised semantic segmentation; Class activation map; Pseudo-label; Level set; Fourier neural operator
College: Faculty of Science and Engineering
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).
Start Page: 111566

Similar Items