A Deep Learning Model for Effective Detection of Early Stages of Alzheimer's Using MRI Images

Abstract

Alzheimer's is a brain disease that typically begins after the age of 40 and manifests with memory loss, cognitive impairment, and a decline in decision-making ability. Alzheimer's is an incurable disease, and its diagnosis usually involves manual evaluation of MRI images and neuropsychological tests. CNN is a deep learning method that is used to diagnose diseases using images. This study proposed a new CNN architecture to detect Alzheimer's. The study can distinguish the early stages of Alzheimer's. The Kaggle MRI image dataset faces the issue of class imbalance. This study addresses this issue by using the Synthetic Minority Over-Sampling Technique (SMOTE) to distribute images evenly among classes. The proposed CNN architecture achieved an accuracy rate of 98.59%.

More by Dr. Adem Tekerek

• Limits of Curtailed Wind Utilization by Ferry-Borne Tourist Electric Vehicles: A Milos Island Case StudyProcesses · 2026
• An Explainable Hybrid Deep Learning Framework Combining Fine-Tuned Convolutional Neural Network Ensembles for Dermoscopic Images of Skin LesionsDermatology Practical & Conceptual · 2026
• Detection of Alzheimer's Disease Using Different Machine Learning Models and Performance Comparison of the ModelsInterdisciplinary Conference on Electrics and Computer · 2026
• Raw URL-Based Phishing Detection Using Domain-Separate Evaluation: A Comparison of CharCNN, Transformer, BiGRU, and LexicalMLPInterdisciplinary Conference on Electrics and Computer · 2026