Multimodal 3D and 2D face authentication approach using extended LBP and statistic local features proposed

In this paper we present a robust face authentication by the fusion of two local methods features and multi-sensors (3D and 2D information). First we fix the rotations of the head in all 3D by Iterative Closest Point (ICP), and then we present our algorithm preprocessing. For feature extraction we use the Extended Local Binary Patterns (E-LBP) and Statistic local features proposed (SLF) based on the calculation of some statistical parameters in a neighborhood of the pixel. For classification we use Principal Component Analysis (PCA) and Enhanced Fisher linear discriminant Model (EFM) methods based for reducing the space followed by the normalized correlation distance. The scores fusion level with a good classifier called Support Vector Machines SVM. The experiments are performed on the CASIA 3D faces database, the experimental results show the promising performance of the proposed system with an EER = 1.16% overall evaluation set and RR = 96.43% in the test set. In our system, all processes are performed automatically.