基于改进的区域生长法的气管与支气管分割

根据气管与支气管的树状特征,提出了一种改进的区域生长法.该方法首先从最上端的CT图像中得到气管初始区域,然后利用气管或支气管在相邻CT图像中的几何特征与灰度特征的相关性对区域生长法进行了改进,其次利用改进的区域生长方法沿着三维空间的树状路径逐步搜索气管与支气管的其他区域,最终得到气管与支气管的全部数据.该方法自动化程度高,不仅适用于气管与支气管分割,还可推广到骨骼、人体血管系统等其他小目标的分割问题.

Segmentation of Trachea and Bronchus Based on Improved Region Growing Method

Considering the tree-like traits of trachea and bronchus,an improved region growing method is presented for their segmentation.Getting the initial region of trachea from the top of CT images,the region growing method is improved taking advantage of the correlation between the geometrical and gray-level characteristics of trachea or bronchus in adjacent CT images.Then,the improved method is applied to the search for other regions of trachea and bronchus step by step along the tree-like paths in 3D space and,...