Advanced Staging of Alzheimer's Disease via ResNet152V2: A High-Precision Approach Using MRI Data

Alzheimer's disease is a neurodegenerative disorder that sustains the loss of neuronal function and connectivity, culminating in cognitive decline and dementia. In such a scenario, it becomes imperative to diagnose and stage Alzheimer's accurately with a view to managing and probably slowing down the progression of the disease. While no cure has been found, early diagnosis by brain imaging with reasonably good accuracy-like through Magnetic Resonance Imaging (MRI)- plays a very essential role in patient management. MRI can point out brain abnormalities linked to different stages of cognitive impairment, such as mild cognitive impairment and Alzheimer's disease. This paper presents a new technique for classifying the stages of Alzheimer's disease from MRI images using the ResNet152V2 architecture, which is very prominent in deep residual learning in a convolutional neural network. The deployed algorithm categorized the MRI images into four different categories: “Mild Demented,” “Moderate Demented,” “Non-Demented,” and “Very Mild Demented.” Model evaluation is done through precision, recall, F1-scores, and accuracy metrics. Results show very high accuracy and robustness in the classification of different stages of dementia, with an overall accuracy of 97.31 % and strong performance for all metrics of classification. Deep learning techniques may more likely add great value to the improvement of diagnostic accuracy and the establishment of a clinical decision-making workflow in Alzheimer's disease management.