Alzheimer’s disease detection from brain MRI: Comparative Study of ResnNet50, VGG19, DenseNet121 Multibranch CNN models

Alzheimer’s disease (AD) is a neurodegenerative condition that results in dementia and severe brain dysfunctions that restrict the ability to perform everyday tasks. Almost 10% of people above the age of 65 suffer from AD according to Alzheimer’s Association. Early detection of AD can help in limiting its extent hence different methods for the early diagnosis of AD have been developed. This research focuses on brain MRIs for their potential use in clinical diagnostics from the ADNI (AD Neuroimaging Initiative) dataset to assess several complex convolutional neural network (CNN) models like ResnNet50, VGG19, DenseNet121 Multibranch CNN models for the diagnosis of AD. Research extensively carried our suggests that multibranch CNN models such as VGG19 and DenseNet have outperformed in classifying non-demented, mild demented, or moderately demented cases with an accuracy of 99.89% and 98.73 % respectively as compared to other traditional approaches for prediction of AD.