Research On Face Recognition Based On PCA

In today’s society, the rapid development of the science and technology has made great progress, at the same time, intelligent products gradually in popularity, so recognition technology more and more widely used in daily life, because face recognition has its unique advantages, so in recent years has been widespread concern and has made great progress, because of face recognition technology includes many fields of professional knowledge, professional knowledge in these areas is crucial for the development of the society, so the research on the technology of face recognition has the significance of the abnormal and has attracted much attention, if mature face recognition is used in anti-terrorism and other fields, undoubtedly it will play a very big help. In this paper, the PCA method is used for face recognition, the main research contents are as follows:First of all, this thesis discusses the background of face recognition technology, as well as its purpose and significance, at the same time, the status quo at home and abroad are discussed, and the current main difficulties to be analyzed. And discussing the process of face recognition is a very important step, namely face pretreatment, because the question of recognition of face images due to differences in the in the acquisition process, subject to the influence of the external environment and its posture, will increase the difficulty of image recognition, to reduce the recognition rate, so that the research of face recognition is particularly important, the pretreatment of several main methods were discussed in detail in this paper.Secondly, introduce the main research methods of this paper, PCA recognition method, because its identification process takes the two-dimensional image into a one-dimensional image dimension, one-dimensional image conversion is very large, the K-L transform can effectively solving this problem, the dimensionality reduction and feature extraction, but this process will increase the feature extraction time, the image recognition time is not ideal, in view of this problem, this paper studied another identification method, 2DPCA identification methods, this method can be directly based on 2D image matrices for feature extraction in image recognition has obvious advantages, to overcome the shortcomings of PCA method, but these two kinds of identification methods have the same shortcomings, is easily affected by external light. In order to overcome the above shortcomings, this paper proposes an improved recognition method, is image enhancement method and 2DPCA recognition method combining, through image enhancement method, the image in the valid information of enhancement, feature extraction, effectively weaken the external light to identify the cause effect.Finally, the three methods experiment in the ORL and AR face database under different conditions. The experimental results show that the improving 2DPCA has the very good recognition effect, compared with the PCA method and 2DPCA method, the improving 2DPCA method has the better recognition rate, and the experients results show that the improving 2DPCA method has very good robustness.