A Four Grade Brain Tumor Classification System Using Deep Neural Network

Abstract

One of the commonly found primary brain tumor is glioma which originate in glial cells. The most commonly found glioma brain tumor is astrocytoma that arises in astrocytes which are star shaped glial cells called astrocytes. The four grade classification of astrocytoma depends upon how fast the tumor grows and its spreading. A novel approach for classifying astrocytoma brain tumor into four grades is proposed. The performance evaluation of the system done is based on statistical measures such as sensitivity, specificity, precision, Mathews Correlation Coefficient, accuracy and FScore. The MRI images are also applied to a Support Vector Machine Classifier and performance is compared. It is found that the performance of the proposed system is very much better than conventional classifiers like SVM.