计算机应用 ›› 2019, Vol. 39 ›› Issue (4): 1201-1207.DOI: 10.11772/j.issn.1001-9081.2018091931

• 虚拟现实与多媒体计算 • 上一篇    下一篇

基于卷积神经网络的左心室超声图像特征点定位

周玉金1,2, 王晓东1, 张力戈1,2, 朱锴1,2, 姚宇1   

  1. 1. 中国科学院 成都计算机应用研究所, 成都 610041;
    2. 中国科学院大学, 北京 100049
  • 收稿日期:2018-09-17 修回日期:2018-10-10 出版日期:2019-04-10 发布日期:2019-04-10
  • 通讯作者: 周玉金
  • 作者简介:周玉金(1993-),女,云南曲靖人,硕士研究生,主要研究方向:深度学习、医学图像处理;王晓东(1973-),男,四川乐山人,研究员,主要研究方向:地理信息系统、物联网;张力戈(1995-),男,山西太原人,博士研究生,主要研究方向:优化算法、深度学习;朱锴(1991-),男,贵州安顺人,博士研究生,主要研究方向:机器学习、深度学习;姚宇(1980-),男,四川宜宾人,研究员,博士,主要研究方向:机器学习、数据挖掘。
  • 基金资助:
    四川省科技厅重点研发项目(2017SZ0010);四川省科技支撑计划项目(2016JZ0035)。

Feature point localization of left ventricular ultrasound image based on convolutional neural network

ZHOU Yujin1,2, WANG Xiaodong1, ZHANG Lige1,2, ZHU Kai1,2, YAO Yu1   

  1. 1. Chengdu Institute of Computer Application, Chinese Academy of Sciences, Chengdu Sichuan 610041, China;
    2. University of Chinese Academy of Sciences, Beijing 100049, China
  • Received:2018-09-17 Revised:2018-10-10 Online:2019-04-10 Published:2019-04-10
  • Supported by:
    This work is partially supported by the Key Research and Development Project of Science and Technology Department of Sichuan Province (2017SZ0010), the Science and Technology Support Project of Sichuan Province (2016JZ0035).

摘要: 针对传统级联卷积神经网络(CNN)在左心室超声图像中定位准确度较低的问题,提出一种融合更快速区域卷积神经网络(Faster-RCNN)模型提取区域的级联卷积神经网络,实现对超声图像中左心室心内膜和心外膜轮廓特征点的定位。首先,采用两级级联的方式改进传统级联卷积神经网络的网络结构,第一级网络利用一个改进的卷积网络粗略定位左心室心内膜和心外膜联合的特征点,第二级网络使用四个改进的卷积网络分别对心内膜特征点和心外膜特征点进行位置微调,之后定位输出左心室心内膜和心外膜联合的轮廓特征点位置;然后,将改进的级联卷积神经网络与目标区域提取融合,即利用Faster-RCNN模型提取包含左心室的目标区域并将目标区域送入改进的级联卷积神经网络;最后,由粗到细对左心室轮廓特征点进行定位。实验结果表明,与传统级联卷积神经网络相比,所提方法在左心室超声图像上的定位效果更好,更逼近真实值,在均方根误差的评价标准下,特征点定位准确度提升了32.6个百分点。

关键词: 超声心动图, 左心室, 特征点定位, 卷积神经网络, 级联卷积神经网络

Abstract: In order to solve the problem that the traditional cascaded Convolutional Neural Network (CNN) has low accuracy of feature point localization in left ventricular ultrasound image, an improved cascaded CNN with region extracted by Faster Region-based CNN (Faster-RCNN) model was proposed to locate the left ventricular endocardial and epicardial feature points in ultrasound images. Firstly, the traditional cascaded CNN was improved by a structure of two-stage cascaded. In the first stage, an improved convolutional network was used to roughly locate the endocardial and epicardial joint feature points. In the second stage, four improved convolutional networks were used to fine-tune the endocardial feature points and the epicardial feature points separately. After that, the positions of joint contour feature points were output. Secondly, the improved cascaded CNN was merged with target region extraction, which means that the target region containing the left ventricle was extracted by the Faster-RCNN model and then was sent into the improved cascaded CNN. Finally, the left ventricular contour feature points were located from coarse to fine. Experimental results show that compared with the traditional cascaded CNN, the proposed method is much more accurate in left ventricle feature point localization, and its prediction points are closer to the actual values. Under the root mean square error evaluation standard, the accuracy of feature point localization is improved by 32.6 percentage points.

Key words: echocardiography, left ventricle, feature points location, Convolutional Neural Network (CNN), cascaded convolutional neural network

中图分类号: