A METHOD FOR COLOR IMAGE SEGMENTATION USING DYNAMIC HISTOGRAM BASED ROUGH-FUZZY C MEANS

摘要

The present invention relates toa method for color image segmentation using Dynamic Histogram based Rough-Fuzzy C-Means (DHRFCM), wherein said method comprises the following steps: (i) Inputting an image of first color model; (ii) Converting image of said first color model to other color models of images than the first color model; (iii)Splitting said image in to four partitions;(iv)Constructing histogram for each partition; (v) Splitting histograms into K clusters; (vi) Finding initial seed point from each histogram; (vii)Reducing seeding points using rough sets; (viii) Applying Fuzzy C-Means (FCM) to each image; (ix) Collecting resultant segmented image; and (x) Retrieving values of three different indices.