An Optimized Framework For Brain Tumor Detection And Classification Using Deep Learning And Texture Analysis

Brain tumors are among the most aggressive diseases, accounting for 85–90% of primary CNS tumors. Annually, about 11,700 people are diagnosed, with a 5-year survival rate of 34% for men and 36% for women. They are classified as benign, malignant, or pituitary tumors. Accurate diagnostics and treatment planning are crucial for improving life expectancy. MRI is the best technique for detecting brain tumors, generating vast image data examined by radiologists. Manual analysis can be errorprone, making AI-driven classification using Deep Learning, specifically CNN and ANN, a reliable alternative. This project aims to enhance MRI image classification accuracy by testing various Deep Learning models and selecting the best-performing one. The goal is to improve diagnosis reliability, track tumor changes, and provide treatment suggestions via cloud and mobile platforms.