Brain Tumor Classification Using Feature Extraction Techniques: A Comparative Study of Local Binary Patterns (LBP) and the Gray Level Co-occurrence Matrix (GLCM) with Random Forest Classification

Brain tumors are atypical proliferations within the brain that may pose a significant threat to life, necessitating prompt and precise diagnosis to enhance patient prognoses and facilitate successful intervention. This work uses MRI data to identify and classify brain tumors using machine learning techniques, including Random Forest and Support Vector Machine (SVM) algorithms. MRI images are denoised, enhanced, and standardized before meaningful features, such as texture patterns, shapes, and statistical characteristics, are extracted. Principal Component Analysis (PCA) is used to decrease dimensionality and improve computational efficiency. Using Random Forest in combination with advanced feature extraction techniques like GLCM, LBP, and PCA, the models achieved an impressive 95% accuracy rate with precision, recall, and F1-scores 0.95, outperforming the SVM model, which achieved a consistent but modest accuracy of 76%. These results highlight the effectiveness of integrating feature engineering and machine learning in complex medical diagnostics. The proposed method is well-suited for real-time clinical applications and automated brain tumor detection pipelines, offering a robust approach to enhancing diagnostic accuracy and reliability.