模糊C均值聚类与多相水平集图割优化相结合的图像分割  

An Image Segmentation Method by Combining Fuzzy C-Means Clustering with Graph Cuts Optimization for Multiphase Level Set Algorithms

作　　者：宋琳[1] 高满屯[1] 王三民[1] 王淑侠[1] 

机构地区：[1]西北工业大学机电学院,陕西西安710072

出　　处：《图学学报》2015年第4期563-569,共7页Journal of Graphics

基　　金：国家自然科学基金资助项目(51105310)

摘　　要：针对在分割多个目标时多相水平集模型对初始轮廓曲线敏感且计算量大的问题,提出采用模糊C均值聚类算法将图像进行粗分割,初始化多相水平集函数,使用图割算法分割出多相结果的方法。该方法能有效减小多相水平集算法对初始轮廓曲线的敏感性,使图割算法在分割图像时更容易分割出理想的目标轮廓;同时,采用图割算法可使水平集函数很快收敛到能量最小值,有效减少计算量,提高计算效率。实验表明该方法具有较好地分割效果和较高地分割效率。Multiphase level set model is sensitive to initial contour curve and has huge computation inthe process of the multiple objects＇ segmentation. A novel Image segmentation method is presentedfor multiphase scenario, which initializes the multiphase level set function by coarse imagesegmentation using fuzzy C-means clustering algorithm and applies graph cuts algorithm to acquiremultiphase output image. The method can effectively reduce the sensitivity of the multiphase level setalgorithm to initial contour and is easier to gain the multiphase output image by graph cuts algorithm.At the same time, the multiphase level set function quickly converge to the minimum energy valuewith small amount of calculation and high computational efficiency using the graph cuts algorithm.The experiments show that this method has better segmentation effect and higher efficiency of imagesegmentation.

关 键 词：模糊C均值聚类 图像分割 图割 多相水平集 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]