Towards Robust Online Domain Adaptive Semantic Segmentation under Adverse Weather Conditions

作     者：Liu, Taorong Xiao, Jing Liao, Liang Lin, Chia-Wen 

作者机构：School of Computer Science Wuhan University China S-lab Nanyang Technological University Singapore Department of Electrical Engineering National Tsing Hua University Taiwan 

出 版 物：《arXiv》 (arXiv)

年 卷 期：2024年

核心收录：

主　　题：Semantic Segmentation 

摘      要：Online Domain Adaptation (OnDA) is designed to handle unforeseeable domain changes at minimal cost that occur during the deployment of the model, lacking clear boundaries between the domain, such as sudden weather events. However, existing OnDA methods that rely solely on the model itself to adapt to the current domain often misidentify ambiguous classes amidst continuous domain shifts and pass on this erroneous knowledge to the next domain. To tackle this, we propose RODASS, a Robust Online Domain Adaptive Semantic Segmentation framework, which dynamically detects domain shifts and adjusts hyper-parameters to minimize training costs and error propagation. Specifically, we introduce the Dynamic Ambiguous Patch Mask (DAP Mask) strategy, which dynamically selects highly disturbed regions and masks these regions, mitigating error accumulation in ambiguous classes and enhancing the model’s robustness against external noise in dynamic natural environments. Additionally, we present the Dynamic Source Class Mix (DSC Mix), a domain-aware mix method that augments target domain scenes with class-level source buffers, reducing the high uncertainty and noisy labels, thereby accelerating adaptation and offering a more efficient solution for online domain adaptation. Our approach outperforms state-of-the-art methods on widely used OnDA benchmarks while maintaining approximately 40 frames per second (FPS). Copyright © 2024, The Authors. All rights reserved.