Application of machine learning for brain tumor diagnosis using magnetic resonance images: a comparative analysis

Abstract

A brain tumor is an abnormal growth of cells that may lead to cancer. MRI scans are the conventional method of diagnosing brain tumors. This paper investigates the potential of machine learning (ML) in interpreting MRI images for brain tumors. The study described applies and evaluates three different methods. The study applied and evaluated three different methods for identifying brain tumors: a self-defined a support vector machine (SVM), a Random forest (RF), and a convolution neural network (CNN). The Bra-TS 2018 dataset is used in this study on MRI brain images containing images of glioma, meningioma, pituitary, and no tumors. Python 3.11 was used for interpreting MRI images for brain tumors. The accuracy of the proposed CNN, RF, and SVM were found to be 99.29%, 99.06%, and 98.36%, respectively. The CNN approach has higher accuracy than innovative techniques.

Keywords

• magnetic resonance imaging
• support vector machine
• random forest
• convolutional neural network
• brain tumor
• machine learning