MCW-Net: Single Image Deraining with Multi-level Connections and Wide Regional Non-local Blocks

29 Sep 2020  ·  Yeachan Park, Myeongho Jeon, Junho Lee, Myungjoo Kang ·

A recent line of convolutional neural network-based works has succeeded in capturing rain streaks. However, difficulties in detailed recovery still remain. In this paper, we present a multi-level connection and wide regional non-local block network (MCW-Net) to properly restore the original background textures in rainy images. Unlike existing encoder-decoder-based image deraining models that improve performance with additional branches, MCW-Net improves performance by maximizing information utilization without additional branches through the following two proposed methods. The first method is a multi-level connection that repeatedly connects multi-level features of the encoder network to the decoder network. Multi-level connection encourages the decoding process to use the feature information of all levels. In multi-level connection, channel-wise attention is considered to learn which level of features is important in the decoding process of the current level. The second method is a wide regional non-local block. As rain streaks primarily exhibit a vertical distribution, we divide the grid of the image into horizontally-wide patches and apply a non-local operation to each region to explore the rich rain-free background information. Experimental results on both synthetic and real-world rainy datasets demonstrate that the proposed model significantly outperforms existing state-of-the-art models. Furthermore, the results of the joint deraining and segmentation experiment prove that our model contributes effectively to other vision tasks.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Single Image Deraining Rain100H MCW-Net PSNR 30.70 # 4
SSIM 0.922 # 2
Single Image Deraining Rain100L MCW-Net PSNR 39.73 # 3
SSIM 0.988 # 1
Single Image Deraining RainCityscapes MCW-Net PSNR 35.82 # 1
SSIM 0.987 # 1

Methods