Detection of Dementia from Brain Tissues Variation in MR Images Using Minimum Cross-Entropy Based Crow Search Algorithm and Structure Tensor Features

Abstract

Dementia causes cognitive dysfunction and deterioration of brain. Alzheimer Disease (AD) and Mild Cognitive Impairment (MCI) are most common forms of dementia. Globally, it is estimated that about 47 million people are affected by dementia. Various researches suggest that AD and MCI share a number of equally severe cognitive deficits, but the pathophysiology has not yet been addressed in a comprehensive way. An attempt is made to observe the prognosis difference in these disorders and to analyze the tissue variation in T1-weighted MR brain images. Samples used in this analysis are obtained from IXI, MIRIAD, and ADNI 2 database. Initially, skull stripping is carried out using Robust Brain Extraction Tool (ROBEX), Brain Extraction Tool (BET), and Brain Surface Extractor (BSE). Further, segmentation of brain tissues is performed using multilevel minimum cross-entropy based Bacteria Foraging Algorithm (BFO) and Crow Search Algorithm (CSA). Various geometric features and Structure Tensor (ST) features are extracted from White Matter (WM), Gray Matter (GM), and Cerebrospinal Fluid (CSF) for normal, MCI, and AD to observe the structural changes. The result shows that ROBEX performs better delineation of brain. Minimum cross-entropy based CSA achieves better segmentation than BFO based on similarity measures and computation time. Further, ST features extracted from the brain tissues are able to show anatomical variation effectively than geometric features. It is identified from the ANOVA test that structure tensor features of GM shows better variation to discriminate normal, MCI, and AD images. Hence, this framework could be used to differentiate normal, MCI, and AD images such as cognitive disorders effectively.