Detecting and Diagnosing Alzheimer's Disease: A Transformer-Based Survey (DAD-T)

Alzheimer's Disease (AD) represents a significant health challenge in today's world. Applying deep learning across various disciplines has demonstrated remarkable potential and has garnered substantial interest. The Transformer model has been shown to have exceptional performance across various computer vision tasks, significantly enhancing its application in medical imaging. Compared to conventional CNN, RNNs, and traditional deep learning algorithms, the Transformer network has demonstrated encouraging results in numerous studies and tasks. This paper reviews multiple works for Mild Cognitive Impairment, early diagnosis, AD and related conditions in the Transformer network over the last half-decade. This paper also presents a comprehensive overview of the data sources, pre-processing, feature extraction, training protocols, and evaluation methods to serve as a resource for future research on AD. Additionally, this survey offers a performance comparison of Transformer models across different datasets, a concise examination of these challenges and suggests potential directions for future research.