SVM for Classification of Brain MRI with K-Means for Detection

Abstract

The segmentation of MR brain images using this method is based on K-means clustering, feature extraction using a discrete wavelet transform (DWT), and feature selection using a grey level co-occurrence matrix (GLCM). For this procedure, we have used a Perfect Radial Basis Function (RBF) - Support Vector Machine (SVM) Classifier. Based on fractions of selectivity and sensitivity, the classifier's performance was measured in terms of accuracy. The proposed classifier was found to be 93% accurate. Additionally, the Histogram methodology was applied in this proposed method in place of randomly choosing the cluster centres. Keywords: K-means clustering, Histon creation, Discrete Wavelet Transform (DWT), GLCM feature selection, and RBF- SVM Classifier.