改进的血管中心路径提取算法  

Improved algorithm for blood vessel centerline extraction

作　　者：吕新荣[1] 高新波[1] 邹华[1] 

机构地区：[1]西安电子科技大学电子工程学院,陕西西安710071

出　　处：《华中科技大学学报（自然科学版）》2009年第9期56-59,共4页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(60771068);教育部长江学者和创新团队支持计划资助项目(IRT0645);国家重点基础研究发展计划资助项目(2006CB705700)

摘　　要：提出一种血管提取速度较快、中心路径提取精度较高的算法.首先,利用八叉树对体数据分块,进行血管的阈值分割和边界距离场计算时,根据每一块体数据的最大最小值与阈值的关系,只处理包含血管体素的体数据块;然后,利用血管体素的梯度值倒数和拉普拉斯变换值作为边界距离场计算的初始值;最后,利用重心法修正初始中心路径使之处于或更接近血管空腔的实际中心路径.选取了2条血管路径进行实验,结果表明本算法能够有效地提高血管提取的速度和中心路径提取的精度.An algorithm for extracting vessels fast and centerlines with higher accuracy was proposed. An octree was introduced to divide the volume dataset into many equal-size blocks. When segmenting blood vessel and computing distance from boundary （DFB）, the minimum and maximum values of each block were compared with the given threshold to determine whether the block should be ignored so that only blocks including blood vessel voxels were processed. The gradient computation and Laplacian transformation were made for all the segmented vessel voxels. The inverse of gradient and Laplacian transformation value of blood vessel voxels were regarded as the initial value of DFB field to improve the accuracy of distance field. The center of gravity was used to fix the initial centerline to make it more close to the real centerline of the vessel cavity. The experimental results show that the proposed algorithm can effectively improve the speed and precision of the centerline extraction.

关 键 词：血管 中心路径 八叉树 边界距离场 梯度 拉普拉斯 重心 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]