Brain Tumor Classification Using Magnetic Resonance Images and Residual Convolutional Neural Networks

This study presents a novel approach for classifying brain tumor MRI images into four categories—glioma, meningioma, no tumor, and pituitary tumor—using a residual convolutional neural network (CNN) enhanced with wavelet denoising and contrast enhancement. The methodology addresses class imbalance and image noise through preprocessing and weighted loss functions. The model, trained on a dataset of ۳,۶۵۹ images, achieved a test accuracy of ۹۱.۶۵% after early stopping at epoch ۵۴. Detailed analysis of precision, recall, and F۱-scores from the confusion matrix highlights robust performance, particularly for the majority classes, with potential for improvement in minority class detection.

Keywords:

Magnetic Resonane Imagin , Convolutional Neural Network , MRI , CNN , Residual Convolutional Neural Network , ResNet