BRAIN TUMOR CLASSIFICATION AND DETECTION USING DEEP LEARNING

Abstract

Brain tumors represent a critical global health challenge, where timely and accurate detection can significantly improve treatment outcomes and survival rates. Manual analysis of MRI scans is prone to delays and misinterpretations due to human error and workload. This project presents a deep learning-based system for automatic classification of brain tumors using Convolutional Neural Networks (CNNs). The model is trained on a diverse MRI dataset sourced from Kaggle and classifies five tumor types: Glioblastoma, Meningioma, Astrocytoma, Pituitary Adenoma, and Medulloblastoma.Our approach employs transfer learning with VGG16 and MobileNet, comparing their performance against traditional machine learning classifiers like Random Forest, Naive Bayes, and Decision Tree. Preprocessing and data augmentation enhance model robustness. The trained models are deployed as a Django-based web application, providing users with real-time predictions, precautionary suggestions, and treatment guidance.