The Research Of Person And ID Card’s Identity Based On Computer Vision

With the rapid development of science and technology, the social activities tend tobe more frequent, the needs of having a stable, fast, accurate personal authenticationand query system become larger and larger; however, currently, face verification hasthe problems such as low efficiency of authentication, poor accuracy and the inadequateuse of information, etc; to this situation, based on the biometric information in thesecond generation ID card, This paper propose one kind of face verification deviceand system based on computer vision.This paper using face recognition for identity verification, face recognition as animportant and complex model in biometrics recognition, with it’s advantages of noninterference,good concealment and environment-friendly, is being widely concerned ofand researched by scholars in the world. This paper will combine the face recognitiontechnology and information of the second generation ID card for face verification. Itsmain contents have three aspects as follow:1.This paper using module2DPCA algorithm for face feature extract, improve theselection of eigenvector and the solution of the average of characteristic matrices inmodule2DPCA algorithm. The features extracted by improved module2DPCAalgorithm can represent face essential feature, the experiment proved our algorithm hasmore effective to face authentication.2. In this paper, to the problem of interference on face authentication by the lightof complex environment, we improve the algorithm by both hardware and software, onthe aspect of hardware, we design a kind of uniform light equipment with the inclinedplane, this device can simulate the light environment when shooting the ID Card; on theaspect of software, we process image by illumination compensation method to furtherreduce the impact of environmental light on the face authentication.3. In face recognition system based on National ID Card II, the interval timebetween the different ID Card holders and the image of ID Card is not fixed, this become a time span problem of face verification. This paper combine the ASM modeland texture modeling，using the characteristic of time insensitive in this two algorithmstime. A block weighted algorithm is designed to reduce the error caused by intervaltime in process of face verification, the experiment proved that our method caneffectively improve the accuracy of face verification.This paper developed the face recognition System based on computer vision,detailed introduced the implementation method of each module and make the relevantexperiments. Experiment results show that this system could quickly and efficientlyverify ID Card Holders and ID card and it has good performance on accuracy androbustness.