Performance evaluation of feature extraction techniques in MR-Brain image classification system

In this paper, we present a MR-Brain image classification system to classify a given MR-brain image as normal or abnormal. This system first employs three feature extraction techniques namely, Gray-Level Co-Occurrence Matrix (GLCM), Local Binary Pattern (LBP) and Histogram of Oriented Gradient (HOG). The obtained feature vector of each technique is passed through a k-Nearest Neighbor (k-NN) classifier. The resulting dissimilarity measure values of the classifiers are combined then by a fusion operator in order to increase the classification accuracy. Two benchmark MR image datasets, Dataset-66 and Dataset-160, have been used to validate the system performance. A cross-validation scheme is adopted to improve the generalization capability of the system. The obtained simulation results are compared with those ones of the existing methods to evaluate the performance of the presented MR-Brain classification system.