Research And Realization Of A Facial Biometric Encryption System

In recent years, biological recognition system has been widely recognized bypeople, but it also has privacy leak problem. At present, in order to ensure thesecurity of the system, researchers have been committed to the encryption systemresearch, so the biometric encryption system arises at this historic moment, and has agood application prospect, but it still in the initial stage, so it have big research space.This paper aim at the problems of equipment, lighting and noise interferenceduring the process of user’s information acquisition, considering the image featureswill be influence authentication system, and The traditional method of Biologicalhash has its limits, for example, the Non-biological characteristics leading theauthentication system, and the feature with high discrimination was difficult toextract from the user information, we put forward some kinds of image preprocessingimprovement schemes to enhance image noise interference resistance ability, afterthat we put forward a new face recognition algorithm based on HAAR wavelettransform and2DPCA. The main research contents include that:First of all, we plan to eliminate the different degree of noise and degradationproblems of image. In this process it mainly from image noise and illuminationcorrection this two directions for processing operation: in the stage of remove the saltand pepper noise, we using the formula of Total Variation method, when the pollutionlevel reaches0.9, it has the good performance; and in the stage of removing lightnoise, we use the iterative illumination for correction.Secondly, put forward the face feature encryption scheme based on the HAARwavelet transform and2DPCA, to reducing the face image information dimensionand structural projection space by HAAR wavelet transform and2DPCA. Thisscheme not only reduce the system running time and improve the system recognitionrate, but also solve the limitation of Non-biological characteristics leading the authentication system, and high degree of differentiate fixed length feature extractionof complexity. At the same time ensure that the key information irreversible systemsafety.Finally, we structures the HAAR wavelet transform and2DPCA face featureencryption system based on MATLAB7.0, and make comprehensive experiment withthis system. At last, make the comparison and analysis with the experiment resultsand show the effectiveness of these methods is correctness and feasibility.