A novel CMC based method for MR! brain image segmentation

Abstract

Segmentation of MRI brain images as a primary step has been significantly used for different medical image analysis applications. In this paper, a new method is proposed based on Combination of Multiple Classifiers (CMC) that is appropriately devised for MRI brain image segmentation. This important category of methods is used for the first time in this area, as our previous work. In this paper, we changed the way that classifiers are cooperating with each other in the ensemble. For achieving this aim, a clustering method is used to divide datasets into different clusters, then a set of classifiers properly cooperates with each other for classifying each test sample. The proposed method is applied on the well-known dataset of IBSR and it can achieved promising results, as compared to the individual classifiers within the ensemble.