MR Brain Image Classification: A Comparative Study on Machine Learning Methods

Abstract

The brain tissue classification from magnetic resonance images provides valuable insight in neurological research study. A significant number of computational methods have been developed for pixel classification of magnetic resonance brain images. Here, we have shown a comparative study of various machine learning methods for this. The results of the classifiers are evaluated through prediction error analysis and several other performance measures. It is noticed from the results that the Support Vector Machine outperformed other classifiers. The superiority of the results is also established through statistical tests called Friedman test.