CD-Buffer: Complementary Dual-Buffer Framework for Test-Time Adaptation in Adverse Weather Object Detection

TL;DR

CD-Buffer introduces a dual-buffer framework that adaptively balances subtractive and additive test-time adaptation strategies based on feature-level domain shift severity, improving object detection in adverse weather conditions.

Contribution

The paper proposes a novel discrepancy-driven coupling mechanism that automatically balances feature removal and refinement strategies without manual tuning.

Findings

Achieves state-of-the-art results on KITTI, Cityscapes, and ACDC datasets.

Effectively handles diverse weather conditions and severity levels.

Demonstrates superior performance across various domain shifts.

Abstract

Test-Time Adaptation (TTA) enables real-time adaptation to domain shifts without off-line 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 the discrepancy-driven coupling: Our framework couples removal and refinement through a unified discrepancy metric, automatically balancing both strategies based on feature-level shift severity. This establishes automatic channel-wise balancing that adapts differentiated treatment to 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.