计算机应用 ›› 2019, Vol. 39 ›› Issue (9): 2731-2737.DOI: 10.11772/j.issn.1001-9081.2019030413

• 虚拟现实与多媒体计算 • 上一篇    下一篇

基于深度可分离卷积和宽残差网络的医学影像超分辨率重建

高媛, 王晓晨, 秦品乐, 王丽芳   

  1. 中北大学 大数据学院, 太原 030051
  • 收稿日期:2019-03-12 修回日期:2019-04-26 出版日期:2019-09-10 发布日期:2019-05-09
  • 通讯作者: 高媛
  • 作者简介:高媛(1972-),女,山西太原人,副教授,硕士,主要研究方向:图像处理、人工智能;王晓晨(1995-),男,山西高平人,硕士研究生,主要研究方向:深度学习、计算机视觉;秦品乐(1978-),男,山西长治人,副教授,博士,主要研究方向:机器视觉、大数据处理;王丽芳(1977-),女,山西长治人,副教授,博士,主要研究方向:机器视觉、大数据处理。

Medical image super-resolution reconstruction based on depthwise separable convolution and wide residual network

GAO Yuan, WANG Xiaochen, QIN Pinle, WANG Lifang   

  1. School of Data Science, North University of China, Taiyuan Shanxi 030051, China
  • Received:2019-03-12 Revised:2019-04-26 Online:2019-09-10 Published:2019-05-09

摘要:

为提高医学影像超分辨率的重建质量,提出了一种基于深度可分离卷积的宽残差超分辨率神经网络算法。首先,利用深度可分离卷积改进网络的残差块,扩宽残差块中卷积层的通道,将更多的特征信息传入了激活函数,使得网络中浅层低级图像特征更容易地传播到高层,提高了医学影像超分辨率的重建质量;然后,采用组归一化的方法训练网络,将卷积层的通道维度划分为组,在每个组内计算归一化的均值和方差,使得网络训练过程更快地收敛,解决了深度可分离卷积扩宽通道数导致网络训练难度增加的问题,同时网络表现出更好的性能。实验结果表明,对比传统的最近邻插值、双三次插值超分辨率算法,以及基于稀疏表达的超分辨率算法,所提算法重建出的医学影像纹理细节更加丰富、视觉效果更加逼真。对比基于卷积神经网络的超分辨率算法,基于宽残差超分辨率神经网络算法和生成对抗网络超分辨率算法,所提算法在峰值信噪比(PSNR)和结构相似性(SSIM)上有显著的提升。

关键词: 超分辨率, 宽残差, 深度可分离卷积, 组归一化, 残差块

Abstract:

In order to improve the quality of medical image super-resolution reconstruction, a wide residual super-resolution neural network algorithm based on depthwise separable convolution was proposed. Firstly, the depthwise separable convolution was used to improve the residual block of the network, widen the channel of the convolution layer in the residual block, and pass more feature information into the activation function, making the shallow low-level image features in the network easier transmitted to the upper level, so that the quality of medical image super-resolution reconstruction was enhanced. Then, the network was trained by group normalization, the channel dimension of the convolutional layer was divided into groups, and the normalized mean and variance were calculated in each group, which made the network training process converge faster, and solved the difficulty of network training because the depthwise separable convolution widens the number of channels. Meanwhile, the network showed better performance. The experimental results show that compared with the traditional nearest neighbor interpolation, bicubic interpolation super-resolution algorithm and the super-resolution algorithm based on sparse expression, the medical image reconstructed by the proposed algorithm has richer texture detail and more realistic visual effects. Compared with the super-resolution algorithm based on convolutional neural network, the super-resolution neural network algorithm based on wide residual and the generative adversarial-network super-resolution algorithm, the proposed algorithm has a significant improvement in PSNR (Peak Signal-to-Noise Ratio) and SSIM (Structural SIMilarity index).

Key words: super resolution, wide residual, depthwise separable convolution, group normalization, residual block

中图分类号: