Abstract
Brain tumors can be detected correctly by MRI images. In this paper, an advanced technique using support vector machine and fuzzy C-means classification has been proposed. Before going to apply the proposed methodology, images should be at high quality to pertain good results. The quality of the image will be enhanced using RGB to gray conversion and followed by employing median filter and binarization techniques. Then fuzzy C-means clustering has been applied to isolate the tumor portion in MRI image of brain. Local Binary Pattern (LBP) has been used to extract features of the brain image, and then SVM classification has been applied to classify the brain MRI images to know whether that tumor is normal or abnormal. The proposed technique provides an accuracy of 94.8%.
