混合聚类彩色图像分割方法研究  

Research for color image segmentation based on hybrid clustering

作　　者：施海滨[1] 周勇[1] 

机构地区：[1]中国矿业大学计算机科学与技术学院,江苏徐州221116

出　　处：《计算机工程与应用》2011年第9期181-184,213,共5页Computer Engineering and Applications

基　　金：国家自然科学基金(No.50674086);江苏省博士后科学基金资助项目(No.0701045B);中国矿业大学科技基金资助项目(No.2007B017)~~

摘　　要：提出了一种基于K-均值算法和EM算法混合聚类的彩色图像分割方法。首先将待分割的RGB彩色图像转化成YUV空间模型,然后将该图像分割成n小块,对每个块的颜色分量用改进的K-均值聚类算法进行聚类分析,最后用EM聚类算法对每个块进行聚类,分割源图像。对K-均值算法和EM算法的初始聚类中心引进了改进算法,加快了算法的收敛速度。并与相似的分割方法进行了比较实验,给出了详细的实验结果与分析。实验表明该方法分割速度快,效果好,具有较高的实用价值。A new color image segmentation algorithm is introduced based on hybrid clustering including K-means algorithm and EM algorithm,which firstly converts RGB color image into YUV-Space model,and then divides the whole image to n-blocks,clusters the color components of each block with K-means improved in this paper,and finally,segments the source image by clustering each block with EM clustering algorithm.In this paper,a new method is proposed to set initial cluster centers in K-means algorithm and EM algorithm,accelerates the convergence speed.The experiment results and the comparison results with similar approach are provided.Experiment results show the proposed algorithm is effective and has high practical value.

关 键 词：图像分割 期望最大化(EM)算法 K-均值算法 YUV颜色空间 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]