Dual-Path Coupled Image Deraining Network via Spatial-Frequency Interaction

TL;DR

This paper introduces DPCNet, a novel image deraining network that combines spatial and frequency domain features using specialized extraction blocks and an adaptive fusion module, leading to superior performance.

Contribution

The paper proposes a dual-path network integrating spatial and frequency information with novel extraction blocks and fusion, addressing limitations of self-attention in capturing high-frequency details.

Findings

Outperforms state-of-the-art deraining methods on six benchmarks

Achieves visually pleasing results with high robustness

Enhances downstream vision task performance

Abstract

Transformers have recently emerged as a significant force in the field of image deraining. Existing image deraining methods utilize extensive research on self-attention. Though showcasing impressive results, they tend to neglect critical frequency information, as self-attention is generally less adept at capturing high-frequency details. To overcome this shortcoming, we have developed an innovative Dual-Path Coupled Deraining Network (DPCNet) that integrates information from both spatial and frequency domains through Spatial Feature Extraction Block (SFEBlock) and Frequency Feature Extraction Block (FFEBlock). We have further introduced an effective Adaptive Fusion Module (AFM) for the dual-path feature aggregation. Extensive experiments on six public deraining benchmarks and downstream vision tasks have demonstrated that our proposed method not only outperforms the existing state-of-the-art deraining method but also achieves visually pleasuring results with excellent robustness on downstream vision tasks.