An Integrated Approach for Face Recognition Using Multi-class SVM

According to recent results, the support vector machine (SVM) classifiers have excellent facial recognition accuracy in the pattern recognition compare to other classification methods. Moreover, this method provides high performance in generalization, processing small samples, and tackling high-dimensional data. Undertaking these advantages of Support Vector Machine (SVM), an efficient novel approach is proposed in this paper, using multi-class SVM to recognize face. In this facial recognition approach, the Histogram of Oriented Gradients (HOG) features extraction method is undertaken to extract facial images feature. Then the oneversus-one SVM method is adopted to accomplish multi-class classification on feature vectors of the facial images. The ORL, the YALE face, and along with Self-created databases are used to perform experiments. The experimental results show the accuracy, and that was above 96 percent recognition rate on both databases, and along with Self-created database.