Background Noise Reduction of Attention Map for Weakly Supervised Semantic Segmentation

Izumi Fujimori, Masaki Oono, Masami Shishibori
4/4/2024
cs.CV

No official code found. Best implementation: YudeWang/SEAM (550 stars, MIT, last commit 2023-07-06).

Abstract

In weakly-supervised semantic segmentation (WSSS) using only image-level class labels, a problem with CNN-based Class Activation Maps (CAM) is that they tend to activate the most discriminative local regions of objects. On the other hand, methods based on Transformers learn global features but suffer from the issue of background noise contamination. This paper focuses on addressing the issue of background noise in attention weights within the existing WSSS method based on Conformer, known as TransCAM. The proposed method successfully reduces background noise, leading to improved accuracy of pseudo labels. Experimental results demonstrate that our model achieves segmentation performance of 70.5% on the PASCAL VOC 2012 validation data, 71.1% on the test data, and 45.9% on MS COCO 2014 data, outperforming TransCAM in terms of segmentation performance.

View on arXivView PDF

Code Implementations(1)

YudeWang/SEAMCommunity72%
MIT

Self-supervised Equivariant Attention Mechanism for Weakly Supervised Semantic Segmentation, CVPR 2020 (Oral)

55095PythonMar 19, 20203 years agoMIT

Similar Papers

FPR: False Positive Rectification for Weakly Supervised Semantic Segmentation

87% match1 repo242026

Leveraging Swin Transformer for Local-to-Global Weakly Supervised Semantic Segmentation

87% match2024

Fine-grained Background Representation for Weakly Supervised Semantic Segmentation

87% match1 repo62024

Spatial Structure Constraints for Weakly Supervised Semantic Segmentation

86% match1 repo122024

SFC: Shared Feature Calibration in Weakly Supervised Semantic Segmentation

86% match1 repo252024

Learning to Detour: Shortcut Mitigating Augmentation for Weakly Supervised Semantic Segmentation

86% match1 repo12024

Cite this paper

@article{fujimori2024background,
  title  = {Background Noise Reduction of Attention Map for Weakly Supervised Semantic Segmentation},
  author = {Izumi Fujimori and Masaki Oono and Masami Shishibori},
  year   = {2024},
  eprint = {2404.03394},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2404.03394v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2404.03394.svg)](https://codeofpaper.com/paper/2404.03394)

Discussion