Recognition and diagnosis of Alzheimer's Disease using T1-weighted magnetic resonance imaging via integrating CNN and Swin vision transformer.

PURPOSE

Alzheimer's disease is a debilitating neurological disorder that requires accurate diagnosis for the most effective therapy and care.

METHODS

This article presents a new vision transformer model specifically created to evaluate magnetic resonance imaging data from the Alzheimer's Disease Neuroimaging Initiative dataset in order to categorize cases of Alzheimer's disease. Contrary to models that rely on convolutional neural networks, the vision transformer has the ability to capture large relationships between far-apart pixels in the images. The suggested architecture has shown exceptional outcomes, as its precision has emphasized its capacity to detect and distinguish significant characteristics from MRI scans, hence enabling the precise classification of Alzheimer's disease subtypes and various stages. The model utilizes both the elements from convolutional neural network and vision transformer models to extract both local and global visual patterns, facilitating the accurate categorization of various Alzheimer's disease classifications. We specifically focus on the term 'dementia in patients with Alzheimer's disease' to describe individuals who have progressed to the dementia stage as a result of AD, distinguishing them from those in earlier stages of the disease.

RESULTS

Precise categorization of Alzheimer's disease has significant therapeutic importance, as it enables timely identification, tailored treatment strategies, disease monitoring, and prognostic assessment.

CONCLUSION

The stated high accuracy indicates that the suggested vision transformer model has the capacity to assist healthcare providers and researchers in generating well-informed and precise evaluations of individuals with Alzheimer's disease.