A new method of face recognition with data field and PCA

In this paper, a new method is proposed on face recognition by integrating data field and PCA(Principal Component Analysis). First, the state of the art is analyzed on PCA face recognition. Second, the method principle is presented. After the features are extracted from facial pictures with data field, faces are recognized by using PCA. Finally, a case is experimented on 400 different faces from ORL (Olivetti Research Lab) face database, for indicating the advantage of the proposed method. The experiments are comparatively done, the results of which are illustrated with the form of tables and figures. The faces are firstly recognized with individual PCA. The result shows PCA has a low recognition rate with few training pictures. Then the faces are redone with the proposed method by integrating PCA and data field. This method just needs a small number of training pictures to get a high recognition rate. So it improves recognition effect of PCA in few training pictures. In practical application, PCA often fails to work because of few training pictures. The new method solves this problem, it has a broad application prospects.