Image reconstruction for face recognition based on PCA and SVM

In this paper, we propose a novel feature extracting method called image reconstruction based on principal component analysis (PCA). The features are used to train support vector machine (SVM) for realize the facial recognition. In this method, a person's images are used and then subspace from each person is obtained. Furthermore, features are obtained by mapping the images to basis images and the features are utilized to reconstruct the images so as to train and test SVMs based on the reconstruction error. The computer simulation illustrates the effectivity of this method on the ORL database.