MRI brain image-based segmentation and classification with optimization using metaheuristic deep learning model in Detection of Alzheimer's disease

Abstract

A brain condition called Alzheimer's disease results in neuronal malfunctions. Dementia and brain function loss brought on by this illness can worsen memory loss, cognitive decline, and behavioural issues in people. Current techniques for diagnosing Alzheimer's using MRI pictures only use specific, targeted subsets of data depending on factors like gender, age, and other characteristics. They also frequently rely on clinical data to help classify the images. The aim of this research is to propose novel technique in MRI brain image segmentation as well as classification utilizing DL technique with metaheuristic model in optimization for Alzheimer's disease detection. In this proposed model the dataset is collected as well as processed for noise removal and segmentation using fuzzy Gaussian C-adaptive histogram equalization. then the segmented image has been classified using support vector convolutional graph transfer VGG-16 learning and optimized using particle grey wolf firefly optimization. The experimental analysis has been carried out for various brain MRI image dataset in terms of Detection accuracy, mini mental state examination (MMSE), weighted average recognition rate (WARR), recall, AUC. Using T1-weighted brain magnetic resonance imaging, an accurate diagnosis of Alzheimer disease was made possible by an autonomous brain segmentation and classification algorithm based on deep learning. Proposed technique MMSE 90%, Detection accuracy 98%, WARR 95%, Recall 94%, AUC 90%.