基于欧氏距离变换的肺实质修补  

LungParenchyma RepairingBased on Euclidean Distance Transformation

作　　者：黄雪艳[1,2] 张莉[1] 

机构地区：[1]中南民族大学生物医学工程学院,武汉430074 [2]深圳市蓝韵网络有限公司,深圳518000

出　　处：《计算机与数字工程》2014年第3期491-494,511,共5页Computer & Digital Engineering

摘　　要：在胸部CT图像的肺实质提取过程中,往往存在诸如过分割等问题,因此有必要对分割出的肺实质图像进行修补,论文提出了一种基于欧氏距离变换的肺实质修补算法。该算法是在初提取肺实质之后,利用欧氏距离变换将肺部边缘的血管和肺结节恢复,从而得到包含病灶区域在内的完整的肺实质。采用该算法对45个病人的胸部CT图像中缺失的肺实质进行了修补,结果显示该算法的正确率达90%以上,证明了该方法的可靠性和有效性。In the lung parenchyma extraction process from chest CT images, there are always such issues as over seg- mentation. It is necessary to repair the lung parenchyma images that have been segmented, so a lung parenchyma patching aI- gorithm based on Euclidean distance transformation is presented. After the initial extraction of the lung parenchyma, the al gorithm uses Euclidean distance transformation to recover blood vessels and pulmonary nodules at the edge of Lung, thereby to get complete lung parenchyma that contains lesion regions. The algorithm has repaired missing lung parenchyma from 45 patients＇ chest CT images. The results show that the algorithm has accuracy more than 90%, which proves the reliability and validity of the method.

关 键 词：肺实质 肺结节 欧氏距离变换 胸部CT 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]