Single Image Deraining via Multi-Scale Degradation-Aware Feature Enhancement

Abstract

Rain streaks with diverse spatial scales and intensities make single image deraining challenging, often causing residual rain artifacts or the loss of fine details. To address this issue, this paper proposes MSDA-Net (Multi-Scale Degradation-Aware Feature Enhancement Network), a multi-scale degradation -aware feature enhancement network for single image deraining. The proposed method extracts rain-related features at multiple spatial scales and adaptively fuses them according to their contributions. A degradation-aware modulation mechanism is further introduced to dynamically regulate feature enhancement strength according to the degradation response of the input image. By integrating multi-scale feature modeling and adaptive enhancement, MSDA-Net improves rain removal while preserving background structures and fine details. Experimental results on several public deraining datasets demonstrate the effectiveness of the proposed method.