Discriminant Gaborfaces and Support Vector Machines Classifier for Face Recognition

Feature extraction, discriminant analysis, and classification rule are three crucial issues for face recognition. This paper presents one method, named GaborfaceSVM, to handle three issues together. For feature extraction, we utilize the Gabor wavelet transform on grey face image to extract Gaborfaces. A Modified Enhanced Fisher Discriminant model is used to reinforce discriminant power of Gaborfaces. During classification process, Support Vector Machines classifier is proposed for robust decision in presence of wide facial and illumination variations. In experiments, the discriminant Gaborfaces approach incorporated with SVMs classifier demonstrates better effectiveness and performance than other methods.