一种基于Canny算子的level-set肺部血管分割算法

Pulmonary Vessel for X-RAY CT Images Segmented Through Canny Level-Set

提出一种应用基于CANNY算子的level-set分割方法,三维分割肺部血管。首先根据肺部解剖结构确定肺部血管不同部位的阈值范围进行对血管初始分割,然后再应用高斯过滤异布扩散对初始血管分割的结果进行平滑去噪和CANNY算子检测出初始图像真正的边缘,最后应用level-set细分初始血管分割的结果,解决由于边界模糊产生肺部血管分割不准确的问题。并提出并描述基于Level-set肺部血管树枝杈点的判定方案。

A method of level-set 3D segmentation for the pulmonary vessel was proposed using CANNY. The Level-set segmentation was based on geometric active contour models instead of parameter active contour models. The method overcame the obscurity of pulmonary vessel＇s boundary by taking advantage of the structure of conforming anatomic arrange of threshold. Further,the initial segmented results preprocessed were applied using anisotropic diffusion and the real border of vessel was detected using CANNY. Finally,the vessel was finely segmented using the level-set method. Thus,the segmentation of pulmonary vessel could be done more accurately. In addition,a judging scheme of branching node of pulmonary vessel tree was proposed on level-set basis.