基于T-snake模型的冠状动脉血管提取和运动跟踪

Coronary Vessel Detection and Motioni Tracking Based on T-snake Model

提出采用拓扑自适应动态轮廓(T-snake)模型对X射线冠状动脉造影图像序列进行二维血管提取和运动跟踪的方法,得到心动周期中冠状动脉的二维形态和运动信息。设计了适用于心血管造影图像的强约束T-snake模型,约束节点沿网格线从一个网格点运动到下一个网格点,通过节点拆分获得拓扑变换能力。针对图像灰度统计特征,设计了使模型能从血管内部的初始位置膨胀变形的能量函数。采用临床采集的X射线冠状动脉造影图像序列对算法进行了验证。

A method to detect and track cardiac vessels from X-ray angiographic equences based on T-snake technique is presented in this paper. Intensive restraint T-snake model is proposed for X-ray coronary artery angiograms. The snake nodes are constrained to move only from one grid vertex to another along grid lines. Topological transformation of the model is implemented through vertex splitting. Proper energy functions,which can drive the model to expand from its initial plan inside the vessel,are designed according to statistical feature of intensities of coronary angiograms. Its effectiveness has been demonstrated on clinical coronary angiograms.