<![CDATA[Early detection of Alzheimer's disease is crucial for effective treatment, and the use of brain MRI images has become a common method for diagnosis. However, many previous studies have faced challenges in addressing class imbalance in their datasets, leading to lower accuracy for minority classes. This study aims to address this issue by using a pretrained CNN architecture, VGG19, combined with the SMOTE method to address class integration and improve classification accuracy. This study contributes by introducing SMOTE to the Alzheimer's MRI image dataset to achieve a more balanced class distribution, which has not been fully explored in previous studies. The evaluation results show that the classification accuracy reaches 95%, higher than previous studies using VGG-19 with an accuracy of 77.66%. These results confirm that the use of VGG19 with SMOTE produces better performance, especially in addressing class representation, which is a key contribution of this study. This research has the potential to be applied in more efficient and accurate automated image-based detection systems, especially for the early diagnosis of Alzheimer's disease.]]>
