A novel approach to detect brain tumour in MRI images using hybrid technique with SVM classifiers

Abstract

The proposed system consists of a hybrid techniques are combining SVM algorithm along with two combined clustering techniques such as k-mean techniques, fuzzy c-mean methods, these all are used to find out the brain tumor. The hybrid techniques are involving image enhancement which is done by contrast improvement and midrange stretch, skull striping is done through double thresholding using morphological operations, segmentation of the image is done through two clustering techniques such as k-means and FCM in which separate analysis is done and also it is also enhanced by combining these k-means and FCM.FCM uses member ship functions to detect real tumor region. The feature extraction is performed by using gray level run length matrix. Finally SVM is helped to classify the image and also grade the location of the tumor is done with sensitivity, specificity, accuracy parameters. GUI program is constructed to test proposed algorithm.