Face Recognition Research Based On Principal Component Analysis

The technology of face recognition is a technology that uses the computer to analyze the image and discriminate identity or recognize status from the worked image. It is a research area spanning several disciplines such as image processing, pattern recognition, computer vision, physiology and psychology. Now it is one of the key issues. However, the fact that the recognition results are very easy to be effected by the variation of the expression, pose and illumination and it is very important to assure the real-time of the recognition system make it difficult for the face recognition (FR) to be practical.The process of FR mainly consists of three parts: Image preprocessing, Feature extraction and Recognition. In the period of preprocessing a method of one dimensional processing is used to extract the edginess image of face. In this method, the image is smoothed using a 1-D Gaussian smoothing filter and 1-D Gaussian differential operator to extract the horizontal and vertical components respectively and then combine two components to obtain the edginess map of the image. The method is better than the approaches based on 2-D operators in the sense that computational time reduced and the smearing of edge information is less.In the period of feature extraction we make use of many techniques including PCA, 2DPCA, (2D)~2PCA, DiagPCA and DiagPCA+2DPCA. Differing from PCA based on image vectors, the methods of 2DPCA, (2D)~2PCA, DiagPCA and DiagPCA+2DPCA are directly based on the image matrices , computationally more efficient than PCA, so these methods can improve the speed of image feature extraction significantly.At last one dimensional processing and the methods of PCA, 2DPCA, (2D)~2PCA, DiagPCA and DiagPCA + 2DPCA are combined to recognize human face. Theexperimental results indicate that the combination of one dimensional processing of images for edginess extraction and the method of 2DPCA, (2D)2PCA and DiagPCA based on the image matrices directly can not only reduce the time of the feature extraction but also improve the robustness to illumination of the recognition system tremendously by means of utilizing of the characteristic of the edginess image.