This paper presents the design and development of a Brain Tumour Detection System developed using
Python and the Flask framework to assist radiologists in the accurate and timely diagnosis of brain
tumours from MRI scans. The system employs deep learning techniques combining a VGG-16 pre-trained
model and transfer learning to analyze MRI images and classify them into four categories — Glioma,
Meningioma, Pituitary, and No Tumour — in real time. Features such as MRI image upload, automated
feature extraction, confidence score generation, and a web-based diagnostic interface are integrated
to enhance usability and clinical reliability. The system is designed with a modular architecture and a
fine-tuned classification pipeline for efficient prediction management. It is observed that the proposed
system improves diagnostic accuracy, reduces radiologist workload, and enhances early detection of
life-threatening neurological conditions.
