Brain Tumor Detection Using Vision Transformer

Authors

  • Umar Rashid The Superior University Lahore
  • Rabia Fareid Department of Computer Science Superior University, Lahore 54600, Pakistan
  • Uswa Zaheer Department of Computer Science Superior University, Lahore 54600, Pakistan

Abstract

Brain tumors exist in a range of different forms, sizes, and features, as well as treatment choices. However, a correct diagnosis can lead to proper treatment. The ViT architecture is able to compete with the classical models by having a better F1-score for the positive class, a very good score for the negative class and also the highest values for the class-weighted metrics with a higher accuracy. The high precision outperforming others at the rate of discovering the brain tumors with other major features. The two common models, the Convolutional Neural Network (CNN), and ViT, which underlines the performance gap between these two modes. The fact, however, is that both algorithms are able and the same time they differ only with brain tumor precise detection, where ViT has better results than the ones obtained by the CNN model. A new Technique, ViT shows a great potential to produce accurate computer-aided Brain tumor diagnosis due to its high accuracy, low error rate and performance consistency when compared to any other metrics used for evaluation. It is through this contrast that this study can recommend the application of ViTs in a variety of medical image processing applications such as cancer detection and even neuron pathology.

Downloads

Published

2024-12-23