Deep Ensemble Learning for Alzheimer’s Disease Classification from Structural MRI Using Fine-Tuned VGG Models

This paper introduces a deep learning model for Alzheimer’s disease (AD) classification from structural MRI scans. It leverages an ensemble of fine-tuned convolutional neural network (CNN) models, VGG16 and VGG19 for discriminative feature extraction and deep neural network classifier for improved classification. The proposed model, trained and evaluated on an augmented Alzheimer MRI dataset, outperforms state-of-the-art approaches, demonstrating improved accuracy, precision, recall, and F1-score. An ablation study further highlights the effectiveness of fine-tuning and ensembling, showing that the fine-tuned ensemble yields the best performance among all evaluated configurations. This study demonstrates the potential of combining transfer learning and ensemble strategies for enhanced AD detection and classification. Future research may explore the use of different CNN architectures or the inclusion of other imaging modalities to further enhance classification performance.