An Intra-Class Mixing Consistency (ICM) framework that enforces prediction consistency between an intra-class mixed image and its original counterpart, demonstrating its effectiveness in mitigating class confusion under challenging environmental conditions is proposed.
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.
Domain adaptive object detection (DAOD) aims to reduce performance degradation due to distribution shifts between training (source) and testing (target) datasets. However, owing to insufficient modeling of fine-grained features and limited instance-level alignment, existing approaches become less effective under comple...
Zhen-Hua Yu, Jia-Hao Zhang, Wen-Chao Zhang et al.· 0 citations
This work proposes REUNION, a hypergraph-guided SFDA framework for object detection in adverse weather conditions, and introduces a group-wise Hyper-InfoNCE objective that optimizes representations at the hyperedge level, enabling semantic information to propagate from confident proposals to corrupted or low-contrast i...
Ji-Cheng Yuan, T. Kieu, Manh Nguyen-Duc et al.· 0 citations
A Content-enhanced Mask TransFormer (CM-Former) is proposed, to enhance the focus of the fundamental component, the mask attention mechanism, in Transformer segmentation models on content information, and to harness the synergistic attributes of these two approaches.
Qi Bi, Shao-Di You, T. Gevers· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.