SMOTE-ADNet: 향상된 CNN 과SMOTE 를 활용한 알츠하이머병 및 초기 단계 정확한 분류

SMOTE-ADNet Leveraging Enhanced CNN and SMOTE for Accurate Classification of Alzheimer's Disease and Early Stages

Alzheimer's disease (AD) is a progressive neurodegenerative disorder characterized by gradual cognitive decline and memory loss, with subtle changes in brain structure that make accurate classification particularly challenging. This study presents SMOTE-ADNet, an innovative Convolutional Neural Network (CNN) model designed to enhance classification performance for Alzheimer's disease by integrating advanced CNN techniques with the Synthetic Minority Over-sampling Technique (SMOTE). The SMOTE-ADNet architecture includes multiple convolutional layers, dropout regularization, and a final dense layer optimized for multi-class classification, aimed at differentiating between five stages of Alzheimer's disease: Alzheimer's disease (AD), Cognitive Normal (CN), Early Mild Cognitive Impairment (EMCI), Late Mild Cognitive Impairment (LMCI), and Mild Cognitive Impairment (MCI). Given the challenge of distinguishing between subtle variations in brain structure during these stages, SMOTE-ADNet effectively balances the dataset using SMOTE and leverages advanced CNN layers to achieve a remarkable accuracy of 98%. This result demonstrates the model's capability to manage the inherent difficulty of classifying subtle structural differences and its potential for improving diagnostic precision and aiding early intervention in Alzheimer's disease.