基于分组卷积和空间注意力机制的单幅图像去雨方法

Single Image De-Rain Method Based on Group Convolution and Spatial Attention Mechanism

为更好地将图像去雨算法应用在户外监控、手机移动终端上,提出一种基于局部空间注意力机制的轻量级卷积神经网络。将图像去雨看作残差学习,既有利于从有雨图中去掉雨滴,又便于模型的训练与优化。深度可分离卷积作为模型提取特征的卷积操作,在不降低模型的性能情况下,显著降低模型的参数量与计算量。局部空间注意力模块利用空洞卷积提供较大的感受野来提取丰富的语义信息,有利于雨滴的检测与去除。在多个公开的数据集上进行对比与测试,证明模型去雨效果较好且速度较快。

In order to better apply the image de-raining algorithm to outdoor monitoring and mobile terminals,a lightweight convolu⁃tional neural network based on the local spatial attention mechanism is proposed.Taking the image to rain as residual learning is not only beneficial for removing raindrops from the raining image,but also for training and optimization of the model.depth-wise convolution,as a convolution operation for extracting features,significantly reduces the amount of parameters and calculations of the model without reducing the performance of the model.The local spatial attention module uses dilate convolution to provide a larger receptive field to extract rich semantic information,which is conducive to the detection and removal of raindrops.Compari⁃son and testing on multiple public data sets prove that the model has better rain removal effect and faster speed.