MR Brain image segmentation for the volumetric measurement of tissues to differentiate Alzheimer’s disease using hybrid algorithm

Abstract

Structural variation in MR brain image is an essential tool to identify Alzheimer’s disease (AD). Gray matter reduction, cortical thickness and the volume of hippocampal which have been very useful to discriminate the severity of AD. These anatomical MRI volumes have been used so far to differentiate AD. In this study, combination of multiple MRI anatomical measures have taken to increase the diagnostic of AD. Patch Image Differential clustering (PIDC) principle is employed to initialize the cluster center. Three various clusters are assigned to segment the Grey Matter(GM), White Matter(WM), Cerebrospinal Fluid(CSF) and hippocampus region and the volume of each segmented region is measured to differentiate AD and Mild Cognitive Imapairment(MCI). Particle Swarm Optimization (PSO) algorithm is also adopted with PIDC to improve the segmentation accuracy. The proposed PIDC with PSO algorithm provided 92% segmentation accuracy and the segmentation results were compared with bench markimages.