Classification of Alzheimer's Disease in Elderly Indian Population using Artificial Intelligence

A major global health concern is Alzheimer's disease, a progressive neurological condition. It mainly affects the elderly people and it affects the memory, thinking skills, and other mental ability skills. The concept presented in this work is the classification of Alzheimer's disease harnessing transfer learning and the feature extraction capabilities of the pre-trained model.4098 Magnetic Resonance Images used for training. Meanwhile, 1023 MRI scan images are used for validation with 4 classes that are Moderate dementated, Mild dementated, Very Mild dementated, and non-dementated. VGG19, EfficientNet, and DenseNet121 pretrained models are used. Among these DenseNet121 performs well in all knowledge transfer, domain adaption, and generalization of the model. The Training accuracy achieved by DenseNet121 was 99.04% with a Validation Accuracy of 75.71% and a Testing Accuracy of 71.04%. This thorough methodology contributes significantly to the continuing attempts to advance diagnostic techniques for neurodegenerative conditions by highlighting the potential of transfer learning in Alzheimer's classification and using DenseNet121's superior performance. The work offers a strong basis for further investigations into improving and fine-tuning models for the classification of Alzheimer's disease