An Efficient Ensemble Approach for Alzheimer's Disease Detection Using an Adaptive Synthetic Technique and Deep...

Area of Science:

• Neurology
• Medical Imaging
• Artificial Intelligence

Background:

• Alzheimer's disease (AD) is an incurable neurodegenerative disorder impacting cognitive function.
• Early AD detection is vital for symptom management and reducing healthcare burdens.
• Traditional machine learning for AD diagnosis from MRI is complex and requires expert feature extraction.

Purpose of the Study:

• To develop an automated, accurate method for early Alzheimer's disease diagnosis using deep learning.
• To address the challenge of imbalanced datasets in Alzheimer's disease MRI datasets.
• To improve diagnostic accuracy and efficiency compared to traditional methods.

Main Methods:

• Utilized an ensemble model combining EfficientNet-B2 and VGG-16 architectures.
• Applied adaptive synthetic oversampling to balance the imbalanced MRI dataset.
• Trained and validated the ensemble model on multi-class and binary-class Alzheimer's disease MRI datasets.

Main Results:

• The proposed ensemble model achieved 97.35% accuracy and 99.64% AUC for multiclass classification.
• For binary classification, the model reached 97.09% accuracy and 99.59% AUC.
• Demonstrated superior performance and efficiency compared to previous Alzheimer's disease diagnostic methods.

Conclusions:

• The EfficientNet-B2 and VGG-16 ensemble model offers a highly accurate and efficient solution for early Alzheimer's disease diagnosis.
• Deep learning, particularly ensemble methods with oversampling, effectively handles imbalanced medical imaging datasets.
• This approach can significantly aid in the early detection of Alzheimer's disease, reducing the strain on medical professionals.