基于优化初始中心的加权K-均值彩色图像聚类算法分析  

Analysis of weighted K⁃mean color image clustering algorithm based on optimized initial center

作　　者：何芳州[1] HE Fangzhou(Criminal Investigation Police University of China,Shenyang 110854,China)

机构地区：[1]中国刑事警察学院,辽宁沈阳110854

出　　处：《现代电子技术》2020年第18期26-29,共4页Modern Electronics Technique

基　　金：国家重点专项(2017YFC0821000);辽宁省自然科学基金计划重点项目(20170540984);中央高校基本科研业务费项目(D2018004)。

摘　　要：超像素作为一种图像预处理技术,在计算机视觉领域中得到了广泛的应用。该文研究的重点就是基于优化初始中心的加权K-均值彩色图像聚类算法的具体应用,针对单板彩色图像,提出一种基于彩色RGB空间通道的K-均值优化初始中心聚类分割方法,并与SLIC进行了分析,得到一种新的SLIC算法(WKK-SLIC算法)。该算法是基于优化加权K-means聚类初始中心点为基础,对彩色图像进行分割处理。在该算法中,通过运用密度敏感相似度对空间像素点密度生成聚类初始中心完成计量,所得聚类结构较稳定且经试验结果表明,在进行图像分割时该算法各方面的性能都要优于其他算法,具有更高的稳定性与准确性。As an image preprocessing technology,the super⁃pixel has been widely used in the field of computer vision.The specific application of the weighted K⁃means color image clustering algorithm based on optimized initial center is the emphases for the research.In allusion to single⁃board color images,a K⁃means optimized initial center clustering segmentation method based on color RGB spatial channel is proposed and analyzed with SLIC,and a new SLIC algorithm(WKK⁃SLIC algorithm)is obtained.The algorithm is based on the optimized initial center points of weighted K⁃means clustering as the basis,so as to conduct the segmentation process for color image.In this algorithm,the clustering initial centers of spatial pixel density are generated to complete the measurement by using density⁃sensitive similarity.The obtained clustering structure is stable,and the experimental results show that the performance of this algorithm is better than that of other algorithms in the image segmentation,and it has higher stability and accuracy.

关 键 词：彩色图像 聚类算法 加权K-均值 优化初始中心 图像分割 试验分析 

分 类 号：TN911.73-34[电子电信—通信与信息系统]