Face image classification using appearance and texture features

Face image classification is a central problem in computer vision research and information retrieval area. Most image classification systems have taken one of two approaches, using either global or local features exclusively. This may be in part due to the difficulty of combining a single global feature vector with a set of local features in a suitable manner. To classify images for versatile applications, an effective algorithm is needed urgently. In this paper, we propose a new texture invariant descriptor to represent global features of an image, and propose a new method which combining local appearance feature with this texture descriptor in face image classification application. Results show the superior performance of these combined method over the hierarchical Bayesian classifier, with a reduction of over 2% in the error rate on a challenging two class dataset from Caltech dataset in face image classification.