Hybrid K -Mean and Refinement Based on Ant for Color Image Clustering
Hybrid K -Mean and Refinement Based on Ant for Color Image Clustering
复制标题
基于 Ant 的混合 K 均值和细化彩色图像聚类
DOI：
10.1007/978-981-10-0135-2_74
复制
发表时间：
2016
期刊：
影响因子：
--
通讯作者：
M. Trivedi
中科院分区：
文献类型：
--
作者：
Lavi Tyagi;M. Trivedi
通过对改进k -均值算法和改进k -均值算法的比较，提出了一种用于彩色图像聚类的混合k -均值算法。首先采用混合算法得到彩色图像聚类结果，然后采用蚁群算法对聚类结果进行细化。在hybridK-mean中，使用modifiedK-mean来解决空聚类形成的问题，使用improvedK-mean来减少每个数据对象到聚类质心之间的距离计算，然后进行细化。实验结果表明，本文提出的混合算法和改进算法能够有效地进行聚类。
By comparing the modifiedK-mean and improvedK-mean algorithm, a hybridK-mean algorithm is proposed for color image clustering. First, hybrid algorithm is applied to get color image clustering result and then ant-based refinement to refine the clustering result. In hybridK-mean, modifiedK-mean is used to solve the problem of empty cluster formation and improvedK-mean to reduce the calculation of distance between each data object and cluster centroid and then refinement will be done. Experimental result shows that proposed hybrid algorithm and refinement effectively and efficiently perform clustering.