Early diagnosis of brain tumors is a major challenge in neuro-oncology. Manual MRI interpretation often suffers from human error and takes too much time. This project introduces a strong, end-to-end deep learning framework that uses the YOLOv11 architecture for accurate classification and pixel-level segmentation of brain tumors. By implementing a Dual-Pipeline "Gatekeeper" system, the model reaches a classification accuracy of 98.7% across four categories: glioma, meningioma, pituitary, and healthy scans. The system is available through a real-time web interface, which speeds up clinical decision-making and offers clear visuals of tumor boundaries.
Shabana A (Sat,) studied this question.