Toward a new face recognition approach using OptimLBP and Learning Classifiers

Face recognition has become a popular biometric technology to fight crime, prevent fraud, and provide security of various human-system interactions. For this reason, researchers have developed several face recognition approaches to minimize the impact of the human factor. In this paper, Optimal Local Binary Pattern (Optim-LBP) features extraction was used to improve face description and preserve features of crucial face components. Afterwards, the 2D Linear Discriminant Analysis (2DLDA) subspace method was used to generate reduced and discriminated face features data and the SVMs Classifiers to perform the classification process. Finally, the effectiveness of the proposed approach was evaluated using the ORL, Yale, and AR face's databases.