融合CNN和Transformer的图像去噪网络

Image Denoising Network Fusing with CNN and Transformer

目前基于深度学习的图像去噪算法无法综合考虑局部和全局的特征信息,进而影响细节处的图像去噪效果,针对该问题,提出了融合CNN和Transformer的图像去噪网络(hybrid CNN and Transformer image denoising network,HCT-Net).首先,提出CNN和Transformer耦合模块(CNN and Transformer coupling block,CTB),构造融合卷积和通道自注意力的双分支结构,缓解单纯依赖Transformer造成的高额计算开销,同时动态分配注意力权重使网络关注重要图像特征.其次,设计自注意力增强卷积模块(self-attention enhanced convolution module,SAConv),采用递进式组合模块和非线性变换,减弱噪声信号干扰,提升在复杂噪声水平下识别局部特征的能力.在6个基准数据集上的实验结果表明,HCT-Net相比当前一些先进的去噪方法具有更好的特征感知能力,能够抑制高频的噪声信号从而恢复图像的边缘和细节信息.

The current image denoising algorithms based on deep learning are unable to consider the local and global feature information comprehensively,which in turn affects the image denoising effect at the details.To address this problem,this study proposes a hybrid CNN and Transformer image denoising network(HCT-Net).First,CNN and Transformer coupling block(CTB)is proposed to construct a two-branch structure that integrates convolution and channel self-attention to alleviate the high computational overhead caused by relying solely on the Transformer.At the same time,the attention weights are dynamically allocated so that the network focuses on important feature information.Secondly,the self-attention enhanced convolution module(SAConv)is designed to adopt the progressive combination of modules and nonlinear transformations to attenuate the noise signal interference and identify local features under complex noise levels.Experimental results on six benchmark datasets show that HCT-Net has better feature perception ability than some current advanced denoising methods and can suppress high-frequency noise signals to recover the edge and detail information of images.