Rethinking Alignment and Uniformity in Unsupervised Image Semantic Segmentation
Abstract
Unsupervised image segmentation aims to match low-level visual features with semantic-level representations without outer supervision. In this paper, we address the critical properties from the view of feature alignments and feature uniformity for UISS models. We also make a comparison between UISS and image-wise representation learning. Based on the analysis, we argue that the existing MI-based methods in UISS suffer from representation collapse. By this, we proposed a robust network called Semantic Attention Network(SAN), in which a new module Semantic Attention(SEAT) is proposed to generate pixel-wise and semantic features dynamically. Experimental results on multiple semantic segmentation benchmarks show that our unsupervised segmentation framework specializes in catching semantic representations, which outperforms all the unpretrained and even several pretrained methods.
Cite
Text
Zhang et al. "Rethinking Alignment and Uniformity in Unsupervised Image Semantic Segmentation." AAAI Conference on Artificial Intelligence, 2023. doi:10.1609/AAAI.V37I9.26325Markdown
[Zhang et al. "Rethinking Alignment and Uniformity in Unsupervised Image Semantic Segmentation." AAAI Conference on Artificial Intelligence, 2023.](https://mlanthology.org/aaai/2023/zhang2023aaai-rethinking-a/) doi:10.1609/AAAI.V37I9.26325BibTeX
@inproceedings{zhang2023aaai-rethinking-a,
title = {{Rethinking Alignment and Uniformity in Unsupervised Image Semantic Segmentation}},
author = {Zhang, Daoan and Li, Chenming and Li, Haoquan and Huang, Wenjian and Huang, Lingyun and Zhang, Jianguo},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {2023},
pages = {11192-11200},
doi = {10.1609/AAAI.V37I9.26325},
url = {https://mlanthology.org/aaai/2023/zhang2023aaai-rethinking-a/}
}