一种基于参数自适应的CT图像肺气道树分割方法  

A Method for Segmenting Lung Airway Tree Based on Adaptive Segmentation Parameter

作　　者：赵鹏[1] 冯焕清[1] 李传富[1,2] 王昌[1] 黄煜峰[1] 

机构地区：[1]中国科学技术大学电子科技系,合肥230027 [2]安徽中医学院第一附属医院影像中心,合肥230031

出　　处：《北京生物医学工程》2009年第5期454-457,共4页Beijing Biomedical Engineering

基　　金：安徽省教育厅重点课题(2006KJ097A);国家自然科学基金(60771007)资助

摘　　要：肺CT图像气道树分割对于肺呼吸功能测定和疾病诊断具有重要意义,但图像噪声和部分容积效应的影响会造成气管分割的泄漏,难于分割出微小的气管。本文利用肺气道树的解剖结构信息,对肺气道树进行分段处理,并提出一种分割参数自适应的方法,动态调整各气管段的分割参数。实验表明,该方法能提高分割的速度和精度,并有效防止泄漏。Segmentation of the tracheo-branchial tree from lung CT images is very helpful in the determination of pulmonary respiratory function and disease diagnosis, however because of the sampling noise and partial volume effects, some tiny branchial bronchus can not be segmented successfully. This paper proposes a method of segmenting airway tree by using anatomic structure knowledge. It segments the airway tree branch by branch and adjusts the segmentation threshold for each branch adaptivelly. And a new searching strategy for the best segmentation parameter in each branch is also designed. The experiment result shows that this method provides a more accurate result with a higher segmenting speed and also prevents the leak effectively.

关 键 词：肺CT图像 气管分割 自适应阈值 区域生长 

分 类 号：R318.04[医药卫生—生物医学工程]