A population based hybrid FCM-PSO algorithm for clustering analysis and segmentation of brain image

Abstract

Fuzzy c-means (FCM) is a well-known unsupervised clustering algorithm based on fuzzy logic and used in many applications. However, it has some disadvantages. One disadvantage of FCM is that, while dealing with complex problems such as medical image data, it is frequently trapped into local minima during execution, which leads the undesired clustering results. Particle swarm optimization (PSO) is a population based metaheuristic optimization algorithm regarded as a global search approach and used in many optimization problems. To overcome the problem in FCM and in order to achieve better results, a hybrid FCM-PSO algorithm has been proposed by combining the excellent features of FCM and PSO algorithms. The experiment has been executed on a triangular dataset and publicly available real brain datasets and compared their results numerically and visually. The obtained experimental results demonstrate the efficacy of the proposed hybrid FCM-PSO algorithm. Friedman’s statistical test is also carried out to demonstrate the statistically significant performance of all discussed algorithms.