基于LVE和ARG算法的冠状动脉分割

为克服传统区域生长方法中容易发生的欠分割和过分割现象,引入局部图像分析技术,设定一系列感兴趣区域(ROI),对冠状动脉的多层螺旋CT(MSCT)图像进行分割。首先应用基于Hessian矩阵的局部血管增强(LVE)滤波,提升图像的对比度;随后采用自适应性区域生长(ARG)算法,并对阈值适时调整。分割后的局部图像经过全局融合得到整体冠脉树。算法综合了图像的局部形状信息和灰度信息,确保了分割结果的准确性和完整性。实验结果表明,算法对左前降支(LAD)、左回旋支(LCX)、对角支(Diag)及右冠状动脉(RCA)均有较好的分割效果。

Segmentation of coronary arteries based on LVE and ARG algorithm

The purpose of this study is to segment coronary arteries in cardiac multislice CT(MSCT) images by sub-dividing the region containing the coronary arteries into a series of small regions of interest(ROI) and analyzing the local shape and intensity feature.First,a Hessian-matrix-based local vessel enhancement(LVE) is performed to enhance the image contrast.Then,the enhanced vascular structures are segmented with an adaptive region growing(ARG) algorithm,which can adjust the threshold in every ROI accordingly...