基于Canny算子的水平集数字虚拟人图像分割算法  

Level Set Segmentation Algorithm for Virtual Human Images Based on Canny Operator

作　　者：秦绪佳[1] 陈新鸿[1] 吴正强[1] 王卫红[1] 

机构地区：[1]浙江工业大学计算机科学与技术学院,杭州310032

出　　处：《系统仿真学报》2010年第7期1674-1677,1682,共5页Journal of System Simulation

基　　金：国家自然科学基金(60673063;60873033);国家"863"高科技计划(2007AA12Z141);国家科技支撑计划(2007BAH11B02);浙江省自然科学基金(Y1080436)

摘　　要：组织器官的分割与提取是医学图像三维重建及可视化能准确表达其相应组织器官的前提。考虑数字虚拟人图像数据的特点,将边缘检测canny算子引入到水平集方法中,提出了一种基于Canny算子的Level Set图像分割算法,推导出了基于Canny算子的Level Set方程的解析表达式,并采用窄带法对该算法进行了数值实现。该算法结合了Canny算子精确定位边界的优点和Level Set图像空间连续演化的思想。实验结果表明,该算法可得到很好的目标分割结果。Segmenting and extracting tissues or organs from medical images are premises of 3D reconstruction and visualization which accurately express the corresponding organs.Considering the features of virtual human images,canny edge detection operator was import into Level set method,and a novel canny operator based Level set algorithm was proposed.The analytic expression of Level set equation was deduced.The algorithm was implemented by numerical using narrow band method.The algorithm was combined with the advantages of canny operator which could orient accurately the edge and the idea that Level set method could evolve the boundary in image defined space continuously.Experiments show that fine results can be obtained by using the algorithm.

关 键 词：CANNY算子 LEVEL SET 窄带法 图像分割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]