3 - Brain MRI image segmentation using nature-inspired Black Hole metaheuristic clustering approach

Abstract

The primary goal of brain image segmentation is to partition the brain image into its constituent homogeneous regions. It is very challenging to classify the pixels into homogeneous regions due to noise and inhomogeneous intensity distribution. This work proposes a framework for MRI brain tissue segmentation based on nature-inspired Black Hole (BH) clustering. The framework consists mainly of two stages. The first stage is to remove nonbrain tissues using a hybrid watershed algorithm (HWA) skull-stripping method. Thereafter, a nature-inspired Black Hole clustering technique is used to segment brain tissues from the skull-stripped image. The proposed framework has been applied on several simulated T1-weighted brain images and the performance can be compared with well-known existing clustering segmentation techniques, such as fuzzy c-means (FCM), k-means (KM), and the gravitational search algorithm (GSA). Experimental results show that the proposed framework performs better than FCM, KM, and GSA on the basis of visual inspection as well as on the basis of JSC, DSC, similarity structure index, and segmentation accuracy evaluation metrics.