基于残差通道注意力网络的医学图像超分辨率重建方法  

Medical-Image Super-Resolution Reconstruction Method Based on Residual Channel Attention Network

作　　者：刘可文[1,2] 马圆[1,2] 熊红霞 严泽军[4] 周志军[5] 刘朝阳[6] 房攀攀[1,2] 李小军 陈亚雷 Liu Kewen;Ma Yuan;Xiong Hongxia;Yan Zejun;Zhou Zhijun;Liu Chaoyang;Fang Panpan;Li Xiaojun;Chen Yalei(School of Information Engineering,Wuhan University of Technology,Wuhan,Hubei 430070,China;Hubei Key Laboratory of Broadband Wireless Communication and Sensor Networks,Wuhan University of Technology,Wuhan,Hubei 430070,China;School of Civil Engineering&Architecture,Wuhan University of Technology,Wuhan,Hubei 430070,China;Department of Urology,Ningbo First Hospital,Key Laboratory of Translational Medicine of Urological Diseases in Ningbo,Ningbo,Zhejiang 315010,China;Department of Urology,the First People′s Hospital of Tianmen,Tianmen,Hubei 431700,China;State Key Laboratory of Magnetic Resonance and Atomic Molecular Physics,Wuhan Institute of Physics and Mathematics,Chinese Academy of Sciences,Wuhan,Hubei 430071,China)

机构地区：[1]武汉理工大学信息工程学院,湖北武汉430070 [2]武汉理工大学宽带无线通信和传感器网络湖北省重点实验室,湖北武汉430070 [3]武汉理工大学土木工程与建筑学院,湖北武汉430070 [4]宁波市第一医院泌尿外科泌尿系疾病转化医学研究宁波市重点实验室,浙江宁波315010 [5]湖北省天门市第一人民医院泌尿外科,湖北天门431700 [6]中国科学院武汉物理与数学研究所波谱与原子分子物理国家重点实验室,湖北武汉430071

出　　处：《激光与光电子学进展》2020年第2期153-161,共9页Laser & Optoelectronics Progress

基　　金：国家重点研发计划(2018YFC0115000)。

摘　　要：针对医学图像超分辨率重建过程中高频信息缺失导致的模糊问题,提出了一种基于残差通道注意力网络的医学图像超分辨率方法。提出的方法在残差网络的基本单元上去除了批规范化层以稳定训练;去掉缩放层、添加通道注意力块,使神经网络更加关注含有丰富高频信息的通道;使用亚像素卷积层进行上采样操作得到最终输出的高分辨率图像。实验结果表明,提出的方法相比主流的图像超分辨率方法在客观评价指标如峰值信噪比和结构相似性上有显著提升,得到的医学图像纹理细节丰富,视觉体验较好。To resolve the fuzzy problem caused by the lack of high-frequency information in the super-resolution reconstruction of medical images,this study proposes a medical-image super-resolution reconstruction method based on a residual channel attention network.The proposed method removes the batch normalization layer from the basic unit of the residual network(ResNet)to stabilize its training.Furthermore,it removes the scaling layer and adds a channel-attention block that focuses the ResNet on channels with abundant high-frequency details.The feature maps are subsampled using a sub-pixel convolution layer,obtaining the final high-resolution images.Experimental results show that the proposed method significantly improves objective evaluation indexes such as the peak signal-to-noise ratio and structural similarity index compared with mainstream image super-resolution methods.The obtained medical images are sufficiently detailed with high visual quality.

关 键 词：图像处理 医学图像处理 图像超分辨率 残差网络 通道注意力机制 亚像素卷积 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]