Computer Face Recognition is an important biological feature recognition technique. Through analyzing face image, it extracts useful features, which can represent face, and identifies person. Computer Face Recognition can be widely used in many fields in the society. Hence many research issues on it have been the hot topic in pattern recognition field.Subspace-based face recognition methods are most popular methods in the current. These methods perform a holistic analysis of the faces, namely project the whole input faces onto a reduced dimensional space in which the recognition is carried out. These methods are different in the optimum criterions, and different statistical properties appear in finding the reduced projection matrix .By analyzing the latest subspace-based face recognition methods, we proposed the improved version of them. Including:1. We apply Partial Least Square algorithm(PLS),Classified Partial Least Square algorithm(CPLS) and Canonical Component Analysis(CCA) algorithm in face recognition. In addition, we point out the essential relationship among the above three algorithms and Fisher, PCA algorithm.2. Researches on kernel-based discriminate vectors on face recognition.Kernel method is a widely used non-linear dimension reduction method. The idea of kernel method is derived from Support Vector Machine, and it differs from traditional method that directly reduces the input space into a lower dimensional space. It projects input space into a very high feature space, and performs original linear methods there. The aim of this conduct is to make the non-classified problem in input space into a classified problem in high dimensional feature space. Because kernel methods only used the inner product of the input sample, the complexity of the computation is not improved.
|