<![CDATA[The rise in global life expectancy has contributed to a rapidly expanding elderly population and a corresponding increase in Alzheimer’s disease cases, highlighting the need for more accurate and objective diagnostic methods. Although MRI is widely used for brain assessment, early-stage Alzheimer’s detection remains challenging because structural differences between disease stages are often subtle and prone to subjective interpretation by clinicians. To address this limitation, this study proposes a custom Convolutional Neural Network (CNN) developed from scratch for classifying Alzheimer’s disease using brain MRI images. Data diversity was enhanced through augmentation comparison strategies, including Albumentations, which achieved 84.8% accuracy; CutMix, which achieved 88.3% accuracy, and a combined Albumentations-CutMix approach, which enabled the base model to achieve 92.1% classification accuracy. Subsequently, a Genetic Algorithm (GA) was applied to optimize key hyperparameters, enabling efficient exploration of the solution space compared to manual tuning and improving model performance to 96.4% accuracy. The optimized model demonstrated improved stability and generalization across all classes, highlighting the capability of the proposed computational framework to function as a reliable tool for supporting the early detection of Alzheimer-related cognitive decline.]]>
