Comparative Analysis of Deep Learning Models for Multiclass Alzheimer’s Disease Classification

Abstract

INTRODUCTION: The terrible neurological condition is known Worldwide; millions of individuals are affected with Alzheimer's disease (AD). Effective treatment and management of AD depend on early detection and a precise diagnosis. An effective method for identifying anatomical and functional abnormalities in the brain linked to AD is magnetic resonance imaging (MRI). OBJECTIVES: However, manual MRI scan interpretation requires a lot of time and is inconsistent between observers. The automated analysis of MRI images for AD identification and diagnosis using deep learning techniques has shown promise. METHODS: In this paper, we present a convolutional neural network (CNN)-based deep learning model for automatically classifying MRI images for Alzheimer's (AD) and a healthy control group. A huge dataset of MRI scans was used to train the CNN, which distinguished between AD and healthy control groups with excellent accuracy. RESULTS: Additionally, we looked into how transfer learning may be used to enhance pre-trained models and boost CNN performance. We discovered that transfer learning considerably increased the model's accuracy and decreased overfitting. Our findings show that MRI scans may be used to precisely detect and diagnose AD utilizing approaches to deep learning and machine learning. CONCLUSION: These techniques may improve the efficiency and accuracy of AD diagnosis and enable early disease identification, resulting in better AD management and therapy.