A Comprehensive Approach to Brain Tumor Classification and Stage Classification Using Neural Networks and K-Means Segmentation

The proposed optimized brain tumor detection method employs a dual-module framework combining advanced image processing, feature extraction, and classification to improve diagnostic accuracy. Brain MRI images are first enhanced using adaptive Wiener filtering for effective noise reduction. Discriminative features are extracted using a Radial Basis Function (RBF) neural network and classified through a Support Vector Machine (SVM) to identify tumor types, including meningioma, glioma, and pituitary tumors. Detected tumors are further analyzed for stage determination—early, intermediate, or advanced— using Convolutional Neural Networks (CNN) and K-means clustering-based segmentation. This integrated approach enables accurate tumor identification and staging, supporting improved treatment planning, timely clinical intervention, and enhanced patient outcomes in medical imaging applications.