Face Recognition Algorithm Based On Principal Component Analysis

Face recognition is a kind of technology which can complete recognizing face by processing and analyzing the face image in the computer, and extracting the representation of face image from the processed image. Face Recognition System(FRS) is comprised of three parts:Preprocessing, Feature extraction and Classification, and the most important technology is feature extraction. At present, there are many technologies of face recognition and each of them has its characteristic。In this article, the popular face recognition technologies are introduced and the method called Principle Component Analysis (PCA) which uses the eigen-face to classify the face images is discussed in detail。The main research work of this article is as follows:(1) The significance and contents of the face recognition is discussed. A brief introduction of FRS is given and the discussion to image preprocessing, feature extracting and classifier designing which are parts of FRS is carried out.(2) The PCA and modular PCA are investigated in detail. The PCA is one of the most successful method of linear differential.But the traditional PCA is influenced by the light condition and the variety of facial expression because it extracts the global feature of the image. The modular PCA is proposed considering this situation. In this method, The image is divided into several modules firstly, then these sub-images' eigenvector is extracted by using PCA. However, the variety of the sub-images is not considered on the traditional modular PCA, an improved modular PCA is proposed and the good performance of this new method is proved by comparing it with the other methods introduced in the paper.(3) The Tow-dimensional PCA(2DPCA) and the modular 2DPCA are discussed. When feature extraction is carried out, the image should be converted to one-dimensional vector firstly, so it makes the vector dimension of image too much high. Because of this, the PCA is extended to 2DPCA and the modular PCA is extended to modular 2DPCA。The covariance matrix dimension of the 2DPCA is far below of the PCA because the former uses the matrix of original image to make the covariance matrix directly. So it can greatly improve the speed and accuracy of feature extraction and reduce the complexity of it. Ultimately, it improve the speed of recognition and the recognition rate.。Considering of ignoring the variety of the sub-images on the traditional modular 2DPCA,an improved modular 2DPCA is proposed. The experiment results indicate that the improved modular 2DPCA is obviously superior to that of traditional modular 2DPCA and other algorithms。(4)The modular (2D)2PCA is proposed. (2D)2PCA is the complete form of 2DPCA.The 2DPCA compresses the image on the direct of line only while the (2D)2PCA does on the both directs. So the eigenvector is smaller and the speed of recognition is faster than the other algorithms. The good performance above the modular 2DPCA is proved by experiments。...