Face recognition using a hybrid SVM–LBP approach and the Indian movie face database

Abstract

Local binary patterns (LBP) are an effective texture descriptor for face recognition. In this work, a LBP-based hybrid system for face recognition is proposed. Thus, the dimensionality of LBP histograms is reduced by using principal component analysis and the classification is performed with support vector machines. The experiments were completed using the challenging Indian Movie Face Database and show that our method achieves high recognition rates while reducing 95% the dimensions of the original LBP histograms. Moreover, our algorithm is compared against some state-of-the-art approaches. The results indicate that our method outperforms other approaches, with accurate face recognition results.