RESNET-50-Based Feature Extraction and Classification Model for Alzheimer’s Disease Detection

Alzheimer disease (AD) is a brain disease that affects any senior people of the country. The burden of care givers can be reduced by detecting the disease at early stage itself. In neuro imaging, deep learning creates new intuitions in finding brain cell variations for detecting various brain diseases such as schizophrenia and Alzheimer’s disease. In this paper, Resnet-50 based deep learning model is constructed for feature extraction and classification to diagnose AD. To do this, we have collected resting state functional magnetic resonance image (rsFMRI) dataset from ADNI 3 database. Initially, the input rsFMRI images are preprocessed and augmented. Then, the dataset is divided into training set and testing set. After that, the RESNET-50-based deep learning feature extraction and classification model is constructed to extract features from the training set and the pre trained model is used to classify AD patients from the test dataset. The experimental results show that this model diagnoses different AD stages with the classification accuracy of 95%, and it also outperforms some recent methods for diagnosing AD.