Face Recognition with Modular Singular Value Decomposition and Radial Basis Probabilistic Neural Networks

摘要

A face recognition algorithm based on singular value decomposition and radial basis probabilistic neural networks is presented. The face images are divided into smaller sub-images and singular value decomposition is applied to each of these sub-images. The larger singular values in each of these sub-images are united a singular value, and these modular singular values of face image are viewed as the eigenvector of face image. These modular singular values are used to train radial basis probabilistic neural networks. The experiment results show that proposed algorithm has a higher recognition rate.