Brain Tumor Classification from MRI Images and Calculation of Tumor Area

Abstract

Medical imaging plays an important role to generate images of mankind for clinical and medical research. In this paper, we are focused on the detection of brain tumor using MRI images, one of the modalities of medical imaging. Basically, tumor is an abandoned growth of tissues in any portion of the human body. Nowadays, automatic detection of brain tumor is the foremost area for research. In this paper, we proposed a system that checks whether the tumor is present or not; if the tumor is present, then classify the tumor. For detection and classification of brain tumor, we have done an experiment on 150 T1-weighted MRI brain images. The supervised classification has applied for classification so that training set is created using texture feature that is extracted with GLCM and DWT methods; for feature selection, principle component analysis has been used. The experimental results show KSVM gives 97% accuracy to classify the brain tumor. We have also calculated the area and volume of the tumor to find stages of the tumor.