Diagnosis of Alzheimer's disease using 2D dynamic magnetic resonance imaging

Zhicheng Zhou, Long Yu, Shengwei Tian, Guangli Xiao

Abstract: Alzheimer’s disease is a degenerative disease of the central nervous system that occurs primarily in old age. Magnetic resonance imaging (MRI) is the most commonly used type of brain medical image in clinical practice to determine the period of Alzheimer’s disease patients. Inspired by clinical practice, we propose using a 2D reparameterization CNN architecture to classify Alzheimer’s disease. The proposed method markedly improves the performance of Alzheimer’s disease classification in less time. By reparametrizing the attention mechanism, the proposed method achieves an AUC of 0.9849 and an ACC of 0.9625.