Hybrid approach for image segmentation using region splitting and clustering techniques

Abstract

Image segmentation plays a significant role in medical diagnosis. In this paper, a hybrid approach of region splitting and clustering segmentation namely KRC technique has been proposed. This KRC algorithm splits an image into 4 regions. The homogeneous pixels in each region are grouped in to clusters according to the intensity values. The clusters in each region are grouped to form new clusters. The different clusters have been merged to form the segmented image. The segmentation results are analyzed based on the quality metrics such as RI (Rand Index), GCE (global consistency error), VOI (variation of information) and processing time using 50 medical images. The Experimental analysis of KRC shows better results based on the quality metrics when compared to existing techniques namely, K-means clustering, Watershed algorithm and region-growing algorithm.