Automatic Segmentation of Natural Color Images in CIE Lab Space using Possibilistic Fuzzy C Means Clustering

Abstract

Clustering is the most significant assignment in image processing. This work performs the segmentation of natural color images in CIELab space based on the Possibilistic fuzzy c means clustering (PFCM).The basic principle of the proposed approach is the segmentation of natural color images based on the two-way approach of hill climbing (HC) and PFCM. In this work, RGB image is transformed into CIELab space for the efficient extraction of the secreted treasure in the images. The combined approach of local optimization search technique, HC and PFCM is applied for the segmentation of synthetic fiber images. This color histogram based technique works on the principle of identification of peaks in the color histogram of the natural color image. The identified peaks are considered as initial seed or clusters. These seeds are then applied to the PFCM to perform the final segmentation. Investigational outcomedemonstrates the competence of thetwo-way approach of HC and PFCMwhich presents the preeminentend result for less complexity color images.