Real-Time Nonlinear Facial Feature Extraction Using Cholesky Decomposition and QR Decomposition for Face Recognition

In this paper, we propose an efficient and effective method for extracting nonlinear discriminative facial features for real-time face recognition tasks. The optimal nonlinear discriminative features of face images are obtained by performing Cholesky decomposition and QR Decomposition only once respectively, which could be implemented using the existing fast algorithms. Since the proposed method does not solve the generalized eigenequation, the high numerical stability is achieved. Moreover, because there is no need to compute the mean of classes and the mean of total samples in the proposed method, the computational complexity is reduced greatly. Thus, the real-time performance for face recognition is guaranteed. The experiments on two standard face databases show that the proposed method can achieve better performance compared with linear facial features extraction method.