Novel face recognition algorithm based on adaptive 3D local binary pattern features and improved Singular Value Decomposition method

Abstract

Face recognition is a kind of important method focused on biological information identification, which is also a research hotspot in the field of pattern recognition and machine vision. In recent years, some pattern recognition researches show that, human visual system uses a lot of visual-based deep information. Therefore, for face recognition in complex environment, we have research focus on depth images based face recognition system, in order to overcome the problem that the 2-D face recognition system is so sensitive to pose, facial expression and illumination changes. It is remarkable that when we apply statistical method to solve the problems of face depth images recognition, we extremely design feature extraction algorithm for specific training sample set. Nevertheless, once these feature extraction algorithms is completed, there will never be any improvement among them. Thus, this situation leads to the poor universality of the feature extraction algorithms, and the effectiveness and stability of the algorithm will be significantly decreased. As the result, the performance of the recognition system is finally affected. In this paper, we focus on the universality problem of feature extraction algorithm and system identification performance, combining feedback learning theory with Neural Network theory and 3-D Local Binary Pattern feature extraction process. We propose a novel face recognition algorithm based on adaptive 3-D Local Binary Pattern and Singular Value Decomposition method. In the process of face recognition, the most important part is facial feature extraction, by the way, Singular Value Decomposition method regards the face images as a matrix, and obtain image features by segmenting face images. The experimental simulation results show that our algorithm has good feature extraction effect and face recognition performance. We also compare our algorithm with other state-of-the-art methodologies and obtain the better effectiveness.