Brain tumor detection using MRI images- a comparative study based on different classifiers

Abstract

The detection of brain tumors is a major challenge in clinical imaging. Integrating machine learning techniques with MRI (Magnetic Resonance Imaging) analysis has been revealed as a powerful and exciting strategy. This study highlights the importance of early detection and precise diagnosis for medical intervention. Machine learning models extract MRI features like texture, shape, and intensity, and train on labeled datasets. The paper discusses the advantages and challenges of this approach, emphasizing data quality, feature engineering, and model selection. It also highlights the potential for continuous improvement in machine learning models. The synergy between machine learning and MRI imaging holds promise for improved patient outcomes and diagnostic processes. This study compares the techniques of ML, DL and Hybrid Learning. This comparative analysis demonstrates that hybrid Learning performs better in identifying BTs on MRI images. Each system produced superior results. Especially, deep CNN+SVM+RBF combined technique yields best performance, with 98.6% accuracy, 98.2% sensitivity, and 98.9% specificity.