ICM: Intra-class Mixing for Domain Adaptation in Adverse Weather

2026-09-23

AI Summary

为解决恶劣天气下无监督领域适应的语义分割问题，提出了一种新的框架ICM，通过保持类内混合图像与其原始图像之间的预测一致性来提高性能。

Abstract

Unsupervised domain adaptation (UDA) for semantic segmentation remains challenging under adverse weather conditions because severe appearance changes enlarge the domain gap and degrade the reliability of pseudo labels in the target domain. To address this problem, we propose an Intra-Class Mixing Consistency (ICM) framework that enforces prediction consistency between an intra-class mixed image and its original counterpart. Unlike previous mixing-based consistency methods that combine regions across different images or domains and may introduce unrealistic semantic inconsistencies, ICM performs mixing within the same image and semantic class, preserving realistic semantic layout for consistency regularization. With ICM, we establish a new state-of-the-art performance for clear-to-adverse-weather unsupervised domain adaptation (UDA) in semantic segmentation. On the Cityscapes $\rightarrow$ ACDC benchmark, our method achieves 75.7\% mIoU, outperforming the previous state of the art by +1.9 pp, demonstrating its effectiveness in mitigating class confusion under challenging environmental conditions. The code is provided in the supplementary material.

Problem

• Unsupervised Domain Adaptation
• Semantic Segmentation
• Adverse Weather Conditions
• Domain Gap
• Pseudo Labels

Innovation

• Intra-Class Mixing
• Consistency Regularization
• Unsupervised Domain Adaptation
• Semantic Segmentation