Face recognition with multi-scale block local ternary patterns

Abstract

In this paper, we propose a novel approach to face recognition, called Multi-scale Block Local Ternary Patterns (MB-LTP), which considers both local and various scale texture information to represent face images. In MB-LTP, we compare average values of sub-regions and use a 3-valued codes method to get the MB-LTP value. The MB-LTP histograms are then extracted and concatenated into a single, spatially enhanced feature vector representing the face image in recognition. We use a nearest neighbor classifier in the computed feature space with Chi square as a dissimilarity measure. MB-LTP code presents several advantages: (1)It is more robust than LBP;(2)it is more discriminative and less sensitive to noise;(3)it encodes not only microstructures but also macrostructures of image patterns. Experiments on ORL and AR databases show that the proposed MB-LTP method significantly outperforms other LBP based face recognition algorithms. © Springer-Verlag 2013.