Adaptive chemical reaction based spatial fuzzy clustering for level set segmentation of medical images

Abstract

Abstract The chemical reaction optimization (CRO), inspired from the interactions of molecules during chemical reactions in reaching a low energy stable state, searches for optimal solution through simulated reactions involving the on-wall ineffective collisions, decomposition, inter-molecular ineffective collision and synthesis. This paper attempts to obtain the global best centroids for spatial fuzzy clustering (SFC) using adaptive CRO (ACRO) with a view to facilitate the level set segmentation for medical images. The approach helps to analyze tumors or unhealthy region in various medical images. The results of medical images of brain, liver, abdomen and eye images are presented to demonstrate the performance.