• Dr. Sanju S. Anand
• Asst. Prof. Ashwini Dsouza
• Vaishnavi Naik
• Varna Poojary
• Shreya P. -Neekha

Brain tumors are serious abnormalities that require careful medical evaluation. Magnetic Resonance Imaging (MRI) is widely used for examining brain structures and abnormalities, but manual interpretation can be time-consuming and requires specialized expertise. This study presents an automated image-classification approach using a Custom Vision Transformer (ViT) for brain MRI images.The proposed system classifies MRI images into four categories: glioma, meningioma, pituitary tumor, and no tumor. The project dataset contains 5,600 training images and 1,600 testing images. Images are converted to RGB format, resized to 224 × 224 pixels, converted to numerical arrays, and normalized before being supplied to the model. The Custom Vision Transformer uses custom Patches and PatchEncoder layers. The trained model achieved a test accuracy of 86.81% and a test loss of 0.8091. The trained model was saved in HDF5 (.h5) format and integrated into a Streamlit application.The application provides a login interface, MRI image upload, predicted tumor category, confidence score, visual tumor/no-tumor indication, and basic information about the predicted category. The system is intended as an academic and research prototype and should not be used as a substitute for professional medical diagnosis.

Brain Tumor, MRI, Vision Transformer, Deep Learning, Image Classification, TensorFlow, Keras, Streamlit.

"A Custom Vision Transformer Approach for Multiclass Brain Tumour Classification Using MRI Images", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.13, Issue 9, page no.a480-a488, September-2026, Available :http://www.jetir.org/papers/JETIR2609058.pdf

"A Custom Vision Transformer Approach for Multiclass Brain Tumour Classification Using MRI Images", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.13, Issue 9, page no. ppa480-a488, September-2026, Available at : http://www.jetir.org/papers/JETIR2609058.pdf