Brain Tumor Detection and Classification using Vision Transformer (ViT)

Abstract

This study uses the Vision Transformer (ViT) architecture to present a sophisticated approach for brain tumor identification and classification. ViT models are assessed based on their capacity to extract global contextual information from MRI images. With an average accuracy of 97%, the refined ViT models surpass traditional CNN-based models on a dataset of brain tumor MRI images classified into glioma, meningioma, pituitary, and no-tumor classes. This study shows how ViT models can increase diagnostic precision, which makes them a good substitute for actual medical imaging applications.