中国组织工程研究 ›› 2010, Vol. 14 ›› Issue (43): 8065-8068.doi: 10.3969/j.issn.1673-8225.2010.43.022

• 骨与关节图像与影像 bone and joint imaging • 上一篇    下一篇

一种基于淘汰粒子群优化聚类算法的心肌超声造影分析系统

杜国庆1,田家玮1,郭延辉○2,宿  阳1,张  萌1   

  1. 1哈尔滨医科大学附属第二医院超声科,心肌缺血机理与诊疗技术省部共建教育部重点实验室(哈尔滨医科大学),黑龙江省哈尔滨市 150086;2犹他州立大学计算机系,美国犹他州 84341
  • 出版日期:2010-10-22 发布日期:2010-10-22
  • 通讯作者: 田家玮,硕士,教授,博士生导师,哈尔滨医科大学附属第二医院超声科,心肌缺血机理与诊疗技术省部共建教育部重点实验室(哈尔滨医科大学),黑龙江省哈尔滨市 150086 duguoqing999@yahoo.com.cn
  • 作者简介:杜国庆☆,男,1972年生,黑龙江省哈尔滨市人,汉族,2008年哈尔滨医科大学毕业,博士,副主任医师,主要从事心血管疾病超声诊断及图像模式识别方面的研究。 duguoqing999@yahoo.com.cn
  • 基金资助:

    教育部高等学校博士学科点专项科研基金(20092307110017)。课题名称“心肌超声造影图像自动分析系统对缺血心肌微循环灌注的研究”;黑龙江省自然科学基金重点项目(ZJY0707-02),课题名称“计算机辅助心肌超声造影结合DSE对缺血心肌功能障碍及机制的研究”;黑龙江省政府博士后资助基金(LRB09-361),课题名称“缺血心肌超声造影图像定量分析软件的设计与实现”;哈尔滨医科大学附属第二医院科研基金(BS2009-20),课题名称“自制超声造影定量系统对缺血心肌微循环灌注的研究”。

A quantitative analysis system for myocardial contrast echocardiography based on eliminating particle swarm optimization clustering algorithm

Du Guo-qing1, Tian Jia-wei1, Guo Yan-hui○2, Su Yang1, Zhang Meng1   

  1. 1 Department of Ultrasound, Second Affiliated Hospital of Harbin Medical University, Key Laboratories of Myocardial Ischemia Mechanism and Treatment (Harbin M edical University), Ministry of Education, Harbin  150086, Heilongjiang Province, China; 2 Department of Computer Science, Utah State University, Logan, UT, 84341, USA
  • Online:2010-10-22 Published:2010-10-22
  • Contact: Tian Jia-wei, Master, Professor, Doctoral supervisor, Department of Ultrasound, Second Affiliated Hospital of Harbin Medical University, Key Laboratories of Myocardial Ischemia Mechanism and Treatment (Harbin Medical University), Ministry of Education, Harbin 150086, Heilongjiang Province, China 150086 duguoqing999@yahoo.com.cn
  • About author:Du Guo-qing☆, Doctor, Associate chief physician, Department of Ultrasound, Second Affiliated Hospital of Harbin Medical University, Key Laboratories of Myocardial Ischemia Mechanism and Treatment (Harbin Medical University), Ministry of Education, Harbin 150086, Heilongjiang Province, China duguoqing999@yahoo.com.cn
  • Supported by:

     the Natural Science Foundation of Heilongjiang Province, No. ZJY0707-02*; Heilongjiang Province Postdoctoral Foundation, No. LRB09-361*; Second Affiliated Hospital of HMU Doctoral Startup Foundation, No.BS2009-20*; Higher Specialized Research Foundation for the Doctoral Program of Ministry of Education No.20092307110017

摘要:

背景:心肌超声造影已成为评价心肌微灌注有效的影像学方法之一,但目前心肌超声造影图像的分析判断多采用目测法或手动描记感兴趣区域,很大程度上依赖观察者的经验,缺乏客观定量分析方法。
目的:探讨新型心肌超声造影图像定量分析系统评价心肌灌注的可行性。
方法:采集家兔心肌超声造影图像,应用基于淘汰粒子群优化聚类(EPSO)算法的计算机图像定量分析系统对造影图像进行自动处理,获得心肌灌注定量参数。
结果与结论: ①图像识别软件自动识别心内膜与心外膜,确定心肌区域。②自动将心肌区域分割为相等的6个区。③自动标记各节段心肌的标化对比剂密度值。④以标化对比剂密度值为依据行彩色编码,获得彩色编码面积百分比。⑤自动生成心肌节段灌注分布直方图。基于淘汰粒子群优化聚类算法的心肌超声图像分析软件,在定量评价心肌微灌注和识别灌注异常方面具有较好的可行性和较高的应用价值。

关键词: 超声心动描记术, 心肌超声造影, 图像处理, 计算机辅助, 心肌灌注, 数字化医学

Abstract:

BACKGROUND: Myocardial contrast echocardiography (MCE) has been used to assess the myocardial capillary perfusion after thrombolytic therapy or percutaneous coronary intervention. The current interpretation of perfusion images by MCE is performed by visual qualitative analysis and drawn manually the region of interests. The accuracy of such analysis depends on the investigators’ experience, and lacks quantitative analysis.
OBJECTIVE: To evaluate the feasibility of a quantitative analysis system for MCE in assessing myocardial perfusion.
METHODS: Rabbit MCE images were collected, and processed automatically by quantitative analysis software based on eliminating particle swarm optimization (EPSO) clustering algorithm to obtain myocardial perfusion parameters.
RESULTS AND CONCLUSION:  ①Quantitative analysis software was used to identify automatically endocardium and epicardium and determine myocardial region; ②Six myocardial regions were dissected automatically; ③Calibrated myocardial contrast intensity (CI) of all segments was calculated automatically; ④Myocardial perfusion of all segments was color-coded according to calibrated CI, and the percentage of color-coded area was obtained; ⑤Myocardial perfusion distribution histogram was generated automatically. MCE image analysis system based on EPSO clustering algorithm in the quantitative assessment of myocardial microperfusion and identification of myocardial perfusion abnormalities has good feasibility and high value.

中图分类号: