Abstract

Alzheimer's disease (AD) is a neurological illness that progresses over time and is a leading global health problem. Effective Alzheimer's disease treatments and better patient outcomes depend on early identification of the illness.Deep learning has become a powerful tool for analysing medical images, including detecting and assessing Alzheimer's disease. This survey analyses existing AD datasets, preprocessing methodologies, and feature extraction methods for deep learning models. We analyse the key contributions and limitations of numerous studies and provide insights into the challenges and opportunities in this domain. The usefulness of several deep learning architectures, including Convolutional Neural Networks (CNNs) and their derivatives, in classifying the MRI scans into mild demented, moderate demented, very mild demented, non demented is examined in our investigation.The paper also examines the role of preprocessing techniques, such as skull alignment and subtraction, in improving the accuracy of deep learning models for AD diagnosis. Additionally, we discuss the potential for further improvement through data augmentation and the collection of more extensive datasets.