Early Diagnosis of Alzheimer's Disease Using YOLOv11 Deep Learning Model in MRI Scans

Alzheimer's Disease (AD) is an anatomical atrophy that attacks important brain regions, specially the memory and it causes the most dangerous dementia form which leads to the death. Millions of peoples in the world suffer from this disorder, and hence, many researches are conducted in order to enhance the early detection and the diagnosis of the disease. Deep learning (DL) models used in computer aided diagnosis (CAD) applications are the most effective nowadays in the classification of Alzheimer's Disease into its different progression stages. YOLO family is among the most recent deep learning architectures, and it has proven its efficiency in many computer vision applications. In this paper, we propose to train magnetic resonance imaging (MRI) data for AD using YOLOv11 pre-trained model for the early detection of AD. The full brain skull was selected as a region of interest (ROI) and the dataset was acquired from Alzheimer's Disease Neuroimaging Initiative (ADNI) database. Four classes of the dementia were selected, which are, ND, MoD, MD and VMD. The best precisions are obtained for ND and VMD stages with 97.6% and 91.1% respectively.