Modified Particle Swarm Optimization Approach applied to MRI Brain Image Clustering

Abstract

his paper proposes a Particle Swarm Optimization Algorithm for Image Clustering, with modified updating mechanisms for particle position. Confinement, wind dispersion and their combination are introduced, compared against standard PSO based clustering and applied on a set of gray-level MRI Brain images. Experimental results show that the PSO-based image clustering with modified updating position performs better than the standard PSO by engendering more compact and well separated clusters.