Research And Realization Of Image Feature Extraction Based On Face Recognition

Face recognition usually refers to an image that already exists in the database to match the image to be detected, then treat the detection of human identity recognition process judgment. Although in recent years the rapid development of technology related to face recognition, but this technology has great interdisciplinary, this technology is currently still relevant in the field of a difficulty.Many image feature extraction algorithm has been shown to have effect. This includes the basic method based on statistical features, such as principal component analysis, linear discriminant analysis, these two methods are the linear methods. In addition to linear algorithm and some nonlinear algorithm also has a good effect, such as various algorithms based on kernel method. Due to the complexity of the image data, so that only the use of algorithms based on extraction effect can be achieved is difficult to identify the requirements for practical use. In recent years, artificial neural network and feature extraction based on the popular learning method favored.The basis of face recognition algorithm based on this specific study on the image feature extraction study and its implementation. Experimental use based on ORL face database, experimental Proof: Consider the sample class information than simple linear discriminant analysis algorithm of dimensionality reduction of principal component analysis algorithm; algorithm, nonlinear methods with respect to the linear feature extraction for various noise It has better robustness. For images with more noise, the better robustness of the algorithm are better able to put into practical use. In the field we rely solely on image feature extraction feature extraction algorithm based on far not meet the requirements of practical application, for more and more effective image feature extraction algorithm research will be carried out in future research work.