Face Recognition System with Distorted Principal Component Analysis and Fuzzy-Gaussian Neural Network

Human face recognition is one of the relative difficult and important issues in the field of pattern recognition , which is significant both in theoretical research and in practical application. This paper present a method of distorted face recognition based on fuzzy-Gaussian neural network (FGNN). At first, this method extracted the eigenface from the distorted images. Based on these extracted facial features, the face drawing is created, which embodies the individual features that the face looks more likely to be so. Secondly, simulating the real transformation of facial images and making corresponding distortions in advance, a series of eigenvector of the distorted facial images are produced. Then the principal components of faces are extracted with Distorted principal component analysis (DPCA), and introduced to train fuzzy neural network. At last, simulation experiment with this method was made. Numerical experimental results on the facial database of Olivetti Research Laboratory (ORL) have shown the effectiveness of the proposed method.