基于分裂合并法和Ncut法的图像分割  

An Image Segmentation Approach Based on Split-Combining Algorithm and Normalized Cut Method

作　　者：邵健萍 陈丽芳[2] 

机构地区：[1]江南大学机械工程学院,江苏无锡214122 [2]江南大学数字媒体学院,江苏无锡214122

出　　处：《江南大学学报（自然科学版）》2011年第5期555-559,共5页Joural of Jiangnan University (Natural Science Edition)　

摘　　要：当灰度图像较大,图像中像素较多时,利用Normalized Cut(Ncut)方法分割生成图的节点数目多,从而给算法求解带来困难,所以提出了利用分裂合并法和Ncut法相结合进行图像分割.用分裂合并法将灰度分割成多个区域,结合区域间的灰度和位置信息,再利用Normalized Cut方法在区域间进行划分,完成图像的分割。实验结果证明该方法有效。When the gray image is larger,and with many pixels,a large number of nodes may be generated in the map divided by the Normalized Cut（Ncut） method,which brings difficulties to the algorithm.Hence the split-Combining Algorithm and Ncut Method are applied to Image Segmentation.Gray area are divided into several parts with the split-Combining Algorithm.Combining with the region between the gray and location information,the image segmentation is completed by using Normalized Cut method to divide between regions.The experimental results show that the method is effective.

关 键 词：图像分割 分裂合并算法 NormalizedCut算法 图割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]