A Novel Clustering Algorithm for Color Image Segmentation using Rough set and Fuzzy Set

Abstract Clustering is a process which partitions a given data set into homogeneous groups based on given features such that similar objects are kept in a group whereas dissimilar Objects are in different groups. This paper describes a proposed algorithm of Dynamic Histogram based Rough-Fuzzy CMeans (DHRFCM), in which identifying initial seed points by constructing dynamic histogram rather than randomly generated initial seed points for identifying similar objects. After that using rough set theory, we can reduce the seed point selection and then apply an un supervised clustering algorithm of Fuzzy C-Means(FCM) to segment the given color image. And also compared the proposed algorithm with various clustering algorithm such as K-means, Fuzzy CMeans, Rough-Fuzzy C-Means. To determine the best clustering algorithm we used various cluster validity indices such as Davies-Bouldin (DB) index, RAND index, Silhouette index and Jaccard index. Experimental results shows that the proposed method perform well and improve the segmentation results.