Wavelet Channel Attention Module with a Fusion Network for Single Image Deraining

TL;DR

This paper introduces a novel CNN with wavelet-based down-sampling and up-sampling, combined with channel attention, to effectively remove rain from single images, improving visibility for computer vision applications.

Contribution

It proposes a wavelet channel attention module with a fusion network that leverages wavelet transforms for better feature extraction and image restoration in deraining tasks.

Findings

Outperforms state-of-the-art deraining methods on synthetic images.

Effectively restores images with diverse rain patterns.

Utilizes wavelet transforms for multi-scale feature learning.

Abstract

Single image deraining is a crucial problem because rain severely degenerates the visibility of images and affects the performance of computer vision tasks like outdoor surveillance systems and intelligent vehicles. In this paper, we propose the new convolutional neural network (CNN) called the wavelet channel attention module with a fusion network. Wavelet transform and the inverse wavelet transform are substituted for down-sampling and up-sampling so feature maps from the wavelet transform and convolutions contain different frequencies and scales. Furthermore, feature maps are integrated by channel attention. Our proposed network learns confidence maps of four sub-band images derived from the wavelet transform of the original images. Finally, the clear image can be well restored via the wavelet reconstruction and fusion of the low-frequency part and high-frequency parts. Several experimental results on synthetic and real images present that the proposed algorithm outperforms state-of-the-art methods.