《计算机应用》唯一官方网站 ›› 2022, Vol. 42 ›› Issue (4): 1308-1316.DOI: 10.11772/j.issn.1001-9081.2021050876

• 多媒体计算与计算机仿真 • 上一篇    

基于卷积神经网络的时频域CT重建算法

李昆鹏1,2, 张鹏程1,2(), 上官宏3, 王燕玲4, 杨婕5, 桂志国1,2   

  1. 1.中北大学 信息与通信工程学院,太原 030051
    2.生物医学成像与影像大数据山西省重点实验室(中北大学),太原 030051
    3.太原科技大学 电子信息工程学院,太原 030024
    4.山西财经大学 信息学院,太原 030006
    5.山西中医药大学 健康服务与管理学院,太原 030024
  • 收稿日期:2021-05-27 修回日期:2021-09-12 接受日期:2021-10-14 发布日期:2021-12-28 出版日期:2022-04-10
  • 通讯作者: 张鹏程
  • 作者简介:李昆鹏(1996—),男,河南信阳人,硕士研究生,主要研究方向:医学图像重建、医学图像处理
    上官宏(1988—),女,山西临汾人,副教授,博士,主要研究方向:模式识别、医学图像处理
    王燕玲(1981—),女,山西吕梁人,讲师,博士,主要研究方向:医学图像处理、数据挖掘
    杨婕(1982—),女,山西太原人,副教授,博士,主要研究方向:精准放射治疗方案优化
    桂志国(1972—),男,天津蓟县人,教授,博士,主要研究方向:信号与信息处理、图像处理与识别、图像重建。
  • 基金资助:
    山西省自然科学基金资助项目(201901D211246);山西省回国留学人员科研资助项目(2016-089);生物医学成像与影像大数据山西省重点实验室基金资助项目(KF2020-60)

Time-frequency domain CT reconstruction algorithm based on convolutional neural network

Kunpeng LI1,2, Pengcheng ZHANG1,2(), Hong SHANGGUAN3, Yanling WANG4, Jie YANG5, Zhiguo GUI1,2   

  1. 1.College of Information and Communication Engineering,North University of China,Taiyuan Shanxi 030051,China
    2.Shanxi Provincial Key Laboratory for Biomedical Imaging and Big Data (North University of China),Taiyuan Shanxi 030051,China
    3.College of Electronic Information Engineering,Taiyuan University of Science and Technology,Taiyuan Shanxi 030024,China
    4.School of Information,Shanxi University of Finance and Economics,Taiyuan Shanxi 030006,China
    5.College of Health Services and Management,Shanxi University of Chinese Medicine,Taiyuan Shanxi 030024,China
  • Received:2021-05-27 Revised:2021-09-12 Accepted:2021-10-14 Online:2021-12-28 Published:2022-04-10
  • Contact: Pengcheng ZHANG
  • About author:LI Kunpeng, born in 1996, M. S. candidate. His research interests include medical image reconstruction, medical image processing.
    SHANGGUAN Hong, born in 1988, Ph. D., associate professor. Her research interests include pattern recognition, medical image processing.
    WANG Yanling, born in 1981, Ph. D., lecturer. Her research interests include medical image processing, data mining.
    YANG Jie, born in 1982, Ph. D., associate professor. Her research interests include optimization of precision radiotherapy.
    GUI Zhiguo, born in 1972, Ph. D., professor. His research interests include signal and information processing, image processing and recognition, image reconstruction.
  • Supported by:
    Natural Science Foundation of Shanxi Province(201901D211246);Scientific Research Funding Project for Returned Overseas Students of Shanxi Province(2016-089);Foundation of Shanxi Provincial Key Laboratory of Biomedical Imaging and Imaging Big Data(KF2020-60)

摘要:

针对采用时域滤波器解析重建后图像存在伪影和图像细节丢失等问题,提出了一种基于卷积神经网络(CNN)的时频域计算机断层扫描(CT)重建算法。首先,在频域中构建了基于卷积神经网络的滤波器网络,实现投影数据的频域滤波;其次,利用反投影操作算子对频域滤波后结果进行域转换得到重建图像;接着,在图像域构建网络对来自反投影层的图像进行处理;最后,在采用最小均方误差损失函数基础上引入多尺度结构相似度损失函数组成复合损失函数,减轻神经网络对结果图像的模糊效应,保留重建图像细节。图像域网络和投影域滤波网络联合作用,最终得到重建结果。在临床数据集上验证了所提算法的有效性,相较于滤波反投影(FBP)算法、全变分(TV)算法及图像域残差编解码CNN(RED-CNN)算法,当投影数目分别为180和90时,所提算法重建结果图像信噪比(PSNR)和结构相似度(SSIM)最高,且归一化均方根误差(NMSE)最小;当投影数目为360时,所提算法仅次于TV算法。实验结果表明,所提算法可以提高CT图像重建图像质量,是一种可行且有效的方法。

关键词: 计算机断层扫描, 数据驱动, 卷积神经网络, 频域滤波, 图像重建

Abstract:

Concerning the problems of artifacts and loss of image details in the analytically reconstructed image by time-domain filters, a new time-frequency domain Computed Tomography (CT) reconstruction algorithm based on Convolutional Neural Network (CNN) was proposed. Firstly, a filter network based on a convolutional neural network was constructed in the frequency domain to achieve the frequency-domain filtering of the projection data. Secondly, the back-projection operator was used to perform domain conversion on the frequency-domain filtered result to obtain a reconstructed image. A network was constructed in the image domain to process the image from the back-projection layer. Finally, a multi-scale structural similarity loss function was introduced on the basis of the minimum mean square error loss function to form a composite loss function, which reduced the blur effect of the neural network on the result image and preserved the details of the reconstructed image. The image domain network and the projection domain filter network worked together to finally get the reconstructed result. The effectiveness of the proposed algorithm was verified on the clinical dataset. Compared with the Filtered Back Projection (FBP) algorithm, the Total Variation (TV) algorithm and the image domain Residual Encoder-Decoder CNN (RED-CNN) algorithm, when the number of projections is respectively 180 and 90, the proposed algorithm achieved the reconstructed result image with highest Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity (SSIM), and the least Normalized Mean Square Error (NMSE).When the number of projections is 360,the proposed algorithm is second only to TV algorithm. The experimental results show that the proposed algorithm can improve the reconstructed image quality of CT image, and it is feasible and effective.

Key words: Computed Tomography (CT), data-driven, Convolutional Neural Network (CNN), frequency domain filtering, image reconstruction

中图分类号: