CD-Buffer：面向恶劣天气目标检测测试时自适应的互补双缓冲框架
CD-Buffer: Complementary Dual-Buffer Framework for Test-Time Adaptation in Adverse Weather Object Detection
提出以特征级域偏移度量为驱动的减法（去除域敏感通道）与加法（轻量特征精炼）双缓冲耦合机制，可即插即用地用于任意检测器在恶劣天气下的在线自适应，为处理不同严重程度分布偏移提供了按通道自适应平衡的通用思路。
原文依据 can we adaptively balance both strategies based on measured feature-level domain shift?
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论测试时自适应（TTA）能够在不进行离线重训练的情况下实时适应域偏移。近期的 TTA 方法主要探索加性途径，即引入轻量模块进行特征精炼。近来，一种去除域敏感通道的减性途径作为替代方向出现。我们观察到这两类范式呈现出互补的有效性模式：减性方法通过去除被污染的特征在严重偏移下表现优异，而加性方法在需要精炼的中等偏移下有效。然而，每种范式仅在有限的偏移严重程度范围内有效，无法泛化到多样的损坏程度。这引出了以下问题：能否根据测得的特征级域偏移自适应地平衡这两种策略？我们提出 CD-Buffer，一种新颖的互补双缓冲框架，其中减性与加性机制在统一差异度量的驱动下沿相反却协同的方向运作。我们的关键创新在于差异驱动的耦合，它通过共享度量统一去除与精炼，并依据特征级偏移严重程度自动平衡两种策略。这建立了通道级自适应平衡，能够对异质的偏移幅度进行差异化处理，而无需人工调参。在 KITTI、Cityscapes 和 ACDC 数据集上的大量实验证明了最优性能，在不同天气条件和严重程度下均持续取得更优结果。
英文原文摘要 1584 字符
Test-Time Adaptation (TTA) enables real-time adaptation to domain shifts without offline retraining. Recent TTA methods have predominantly explored additive approaches that introduce lightweight modules for feature refinement. Recently, a subtractive approach that removes domain-sensitive channels has emerged as an alternative direction. We observe that these paradigms exhibit complementary effectiveness patterns: subtractive methods excel under severe shifts by removing corrupted features, while additive methods are effective under moderate shifts requiring refinement. However, each paradigm operates effectively only within limited shift severity ranges, failing to generalize across diverse corruption levels. This leads to the following question: can we adaptively balance both strategies based on measured feature-level domain shift? We propose CD-Buffer, a novel complementary dual-buffer framework where subtractive and additive mechanisms operate in opposite yet coordinated directions, driven by a unified discrepancy metric. Our key innovation lies in discrepancy-driven coupling, which unifies removal and refinement through a shared metric and automatically balances both strategies based on feature-level shift severity. This establishes channel-wise adaptive balancing that enables differentiated treatment for heterogeneous shift magnitudes without manual tuning. Extensive experiments on KITTI, Cityscapes, and ACDC datasets demonstrate state-of-the-art performance, consistently achieving superior results across diverse weather conditions and severity levels.