[1]郭翌,吴香奕,吴茜,等.基于循环一致生成对抗网络的多模态影像刚性配准[J].中国医学物理学杂志,2021,38(2):198-203.[doi:DOI:10.3969/j.issn.1005-202X.2021.02.013]
　GUO Yi,WU Xiangyi,WU Qian,et al.Rigid registration of multimodal images based on CycleGAN[J].Chinese Journal of Medical Physics,2021,38(2):198-203.[doi:DOI:10.3969/j.issn.1005-202X.2021.02.013]
点击复制

基于循环一致生成对抗网络的多模态影像刚性配准()
分享到：

《中国医学物理学杂志》[ISSN:1005-202X/CN:44-1351/R]

卷:
38卷
期数:
2021年第2期
页码:
198-203
栏目:
医学影像物理
出版日期:
2021-02-02

文章信息/Info

Title:
Rigid registration of multimodal images based on CycleGAN
文章编号:
1005-202X（2021）02-0198-06
作者:
郭翌1; 吴香奕1; 吴茜2; 陈志1; 徐榭1; 裴曦1
1.中国科学技术大学物理学院工程与应用物理系， 安徽 合肥 230026； 2.安徽医科大学， 安徽 合肥 230032
Author(s):
GUO Yi1;  WU Xiangyi1;  WU Qian2;  CHEN Zhi1;  XU Xie1;  PEI Xi1
1. Department of Engineering and Applied Physics, University of Science and Technology of China, Hefei 230026, China 2. Anhui Medical University, Hefei 230032, China
关键词:
医学影像; 刚性配准; 多模态; 循环一致; 生成对抗网络
Keywords:
Keywords: medical images rigid registration multimodal cycle-consistent generative adversarial networks
分类号:
R318
DOI:
DOI:10.3969/j.issn.1005-202X.2021.02.013
文献标志码:
A
摘要:
目的：采用循环一致生成对抗网络（CycleGAN）方法在保证医学影像刚性配准精度的同时，降低图像配准的时间以及训练数据获取难度。方法：首先对训练数据进行标准化与归一化，同时对图像进行重采样与剪裁，去除多余的空气部分。其次采用阈值法与扫描线法获取图像的外轮廓信息，基于CycleGAN建立两个生成器与两个判别器，生成器输入配准图像对和输出配准结果，判别器输入配准图像对和输出配准程度。在原始CycleGAN损失函数基础上，增加轮廓损失项，以约束网络训练方向，提高收敛速度。结果：选取75例腹部病例，其中65例作为训练数据集，10例作为测试数据集，配准结果与配准软件Elastix对比。计算测试图像集外轮廓Dice系数，配准前图像对的平均Dice系数为0.858，Elastix配准后的平均Dice系数为0.926，本方法配准后的平均Dice系数为0.925。配准时间上Elastix的平均配准时间为12.1 s，本研究方法的平均配准时间为0.04 s，加速比达到302。结论：本方法在保证图像配准精度的同时极大降低了图像配准所需的时间，提高了配准流程工作效率。除此之外，与其他深度学习网络相比，本方法不需要真实配准结果以及传统相似性测度。
Abstract:
Abstract: Objective To reduce the time of image registration while ensuring the rigid registration accuracy of medical images. Methods Firstly, the training data were standardized and normalized. At the same time, the images were re-sampled and clipped to remove the redundant air. Secondly, threshold method and scanning line method were used to obtain the outline information of the images. Based on cycle-consistent generative adversarial networks (CycleGAN), two generator networks and two discriminator networks were established. The generator networks received multimodal image pairs and output the registration results. The discriminator networks received registration results and output registration accuracy. To improve the convergence speed of CycleGAN, the contour loss was added to restrict the training direction of the network. Results 75 abdominal cases were selected, of which 65 cases were used as training data set and 10 cases were used as test data set. Results of registration were compared with those of the registration software Elastix. The average Dice coefficient of the image pairs before registration was 0.858, that after Elastix registration was 0.926, and that after CycleGAN registration was 0.925. The average registration time of Elastix is 12.1 s. However, CycleGAN took 0.04 s to finish image registration, and the acceleration ratio was 302. Conclusion The proposed method not only ensures the accuracy of image registration, but also greatly reduces the time required for image registration and improves the efficiency of the registration process. In addition, compared with other deep learning methods, this method does not need real registration results and similarity measures.

相似文献/References:

[1]王弈,李传富.人工智能方法在医学图像处理中的研究新进展[J].中国医学物理学杂志,2013,30(03):4138.[doi:10.3969/j.issn.1005-202X.2013.03.013]
[2]朱翔宇,葛中芹,张冰清,等.基于图像处理的医学影像处理平台系统设计[J].中国医学物理学杂志,2017,34(4):388.[doi:DOI:10.3969/j.issn.1005-202X.2017.04.014]
[3]张明,吴凉,张文莉. 一种符合医学数字影像与通讯标准的医学影像管理系统的设计[J].中国医学物理学杂志,2018,35(3):281.[doi:DOI:10.3969/j.issn.1005-202X.2018.03.007]
　ZHANG Ming,WU Liang,ZHANG Wenli. Design of a medical image management system conforming to DICOM standard[J].Chinese Journal of Medical Physics,2018,35(2):281.[doi:DOI:10.3969/j.issn.1005-202X.2018.03.007]
[4]皮一飞,吴茜,裴曦,等.基于掩膜优化的多模态医学图像刚性配准[J].中国医学物理学杂志,2018,35(9):1022.[doi:10.3969/j.issn.1005-202X.2018.09.006]
　PI Yifei,WU Qian,PEI Xi,et al.Rigid registration of multimodal medical images based on mask optimization[J].Chinese Journal of Medical Physics,2018,35(2):1022.[doi:10.3969/j.issn.1005-202X.2018.09.006]
[5]李双双,侯震,刘娟,等.影像组学分析与建模工具综述[J].中国医学物理学杂志,2018,35(9):1043.[doi:10.3969/j.issn.1005-202X.2018.09.010]
　LI Shuangshuang,HOU Zhen,LIU Juan,et al.Review of radiomic analysis and modeling tools[J].Chinese Journal of Medical Physics,2018,35(2):1043.[doi:10.3969/j.issn.1005-202X.2018.09.010]
[6]胡晓伟,孙新臣,李金凯,等. 自动勾画轮廓软件在头颈部锥形束CT中的应用[J].中国医学物理学杂志,2019,36(3):254.[doi:DOI:10.3969/j.issn.1005-202X.2019.03.002]
　HU Xiaowei,SUN Xinchen,LI Jinkai,et al. Application of automatic segmentation software in head and neck cone beam CT[J].Chinese Journal of Medical Physics,2019,36(2):254.[doi:DOI:10.3969/j.issn.1005-202X.2019.03.002]
[7]赵瑞斌,王安琪,吴艳茹,等. X线机工作特性模拟实验装置设计与研制[J].中国医学物理学杂志,2019,36(9):1063.[doi:DOI:10.3969/j.issn.1005-202X.2019.09.013]
　ZHAO Ruibin,WANG Anqi,WU Yanru,et al. Design and development of an experimental device to simulate the working characteristics of X-ray machine[J].Chinese Journal of Medical Physics,2019,36(2):1063.[doi:DOI:10.3969/j.issn.1005-202X.2019.09.013]
[8]张富利,王雅棣,王秋生.放射治疗中医学图像自动分割的研究进展[J].中国医学物理学杂志,2021,38(9):1108.[doi:10.3969/j.issn.1005-202X.2021.09.011]
　ZHANG Fuli,WANG Yadi,WANG Qiusheng.Advances in automatic segmentation of medical images in radiotherapy[J].Chinese Journal of Medical Physics,2021,38(2):1108.[doi:10.3969/j.issn.1005-202X.2021.09.011]
[9]王霄,朱恩照,艾自胜.卷积神经网络的原理及其在医学影像诊断中的应用[J].中国医学物理学杂志,2022,39(12):1485.[doi:DOI:10.3969/j.issn.1005-202X.2022.12.005]
　WANG Xiao,ZHU Enzhao,AI Zisheng.Principle of convolutional neural network and its applications in medical imaging diagnosis[J].Chinese Journal of Medical Physics,2022,39(2):1485.[doi:DOI:10.3969/j.issn.1005-202X.2022.12.005]
[10]宫宁,王瑜,王逸夫,等.基于反距离加权插值和改进采样点数据合成的医学影像三维重建[J].中国医学物理学杂志,2022,39(12):1510.[doi:DOI:10.3969/j.issn.1005-202X.2022.12.009]
　GONG Ning,WANG Yu,WANG Yifu,et al.Three-dimensional medical image reconstruction using inverse distance weight and improved data synthesis of sampling points[J].Chinese Journal of Medical Physics,2022,39(2):1510.[doi:DOI:10.3969/j.issn.1005-202X.2022.12.009]

备注/Memo

备注/Memo:
【收稿日期】2020-09-13 【基金项目】安徽省自然科学基金（1908085MA27）；国家自然科学基金（11575180）；安徽省重点研究与开发计划（1804a09020039）；安徽省高校自然科学研究项目（KJ2019A0240） 【作者简介】郭翌，在读博士，研究方向：医学影像配准与可视化，E-mail: gyi@mail.ustc.edu.cn 【通信作者】裴曦，E-mail: xpei@ustc.edu.cn
更新日期/Last Update: 2021-02-04