基于差分进化算法的图像聚类研究  

Research on image cluster based on differential evolutionary algorithm

作　　者：贾紫娟[1] 杨淑莹[1] 王光彪[1] 

机构地区：[1]天津理工大学智能计算及软件新技术重点实验室,天津300384

出　　处：《天津师范大学学报（自然科学版）》2012年第2期53-55,共3页Journal of Tianjin Normal University：Natural Science Edition

基　　金：国家自然基金资助项目(61001174);天津市高校发展基金资助项目(20071308)

摘　　要：将差分进化算法应用于图像聚类问题,对问题进行实数编码,采用群体智能模式实现问题解的搜索.利用差分进化算法的差分变异操作和群体分布特性有效提高算法的搜索能力,采用贪婪选择操作和竞争生存策略实现群体内个体之间的相互合作与竞争,降低了进化操作的复杂性,并通过仿真实验证明了该算法的有效性.Differential evolutionary algorithm is applied to image clustering problem.The problem is firstly encoded using real number,and swarm intelligent model is used for achieving the search for the solution.Through taking the best of features of the group distribution,the search capability of the algorithm is improved by using the operation of differential variation of the differential evolutionary algorithm.Mutual cooperation and competition among individuals within the group is achieved by adopting greed selection and survival strategies,and the complexity of the evolutionary operation is reduced.The simulation experiments show that the proposed algorithm for image cluster is correct and effective.

关 键 词：差分进化算法 群体智能 图像聚类 优化问题 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]