一种分割冠状动脉X射线造影图像的有效方法  

An effective approach for coronary artery segmentation from X-ray angiographic image

作　　者：周辉[1] 何家驹[1] 杨媛[2] 向征[1] 谢斓[1] 李绍林[1] 周寿军[1] 

机构地区：[1]中国人民解放军第四五八医院,广东广州510602 [2]广州军区联勤部沿江东路干部休养所卫生所,广东广州510110

出　　处：《生物医学工程与临床》2013年第4期312-315,共4页Biomedical Engineering and Clinical Medicine

基　　金：国家自然科学基金资助项目(61179020;31000450;60902103)

摘　　要：目的提出一种简单有效的方法进行冠状动脉X射线造影图像的分割。方法基于Hessian矩阵的多尺度滤波和区域增长等算法,其中多尺度滤波用来增强造影图像中的血管,然后利用多种子点区域增长算法从增强后的图像中提取冠状动脉树。结果该方法对于造影图像中血管状结构非常敏感,能够清晰提取出冠状动脉树中较细小的末梢,并能有效抑制噪声。结论该方法适合于分割冠状动脉造影,适用于冠状动脉造影的精确量化分析。Objective To propose a simple and powerful method for segmentation of the coronary artery in angiogram images.Methods The Hessian matrix multi-scale filtering was used to enhance vessel in angiographic images and the multi-seed region-growing algorithm was subsequently used to extract the coronary artery tree from the enhanced images.Results The proposed method was sensitive to vessel-like structures in angiograms and inhibited noise,both the coronary artery trees and smaller distal vessels could extracted clearly.Conclusion It is demonstrated that the method is suitable for the segmentation of coronary angiography and further accurate quantitative analysis in coronary angiography.

关 键 词：冠状动脉造影 HESSIAN矩阵 区域增长 血管分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]