The Research And Implementation Of Authentication Based On Combination Of Face And Voiceprint

As the technology of information and intelligent obtaining the popularization,the protection of information security and privacy calls more and more attention in life.Authentication is necessary in many occasions and its requirements are changing all the time.How to improve the security,accuracy and practicality of the authentication has become a problem can't be ignored in the verification progress.To solve these problems,this paper makes the following research:(1)It analyzes the shortcomings and limitations of existing traditional and single biometric authentication and puts forward the solution based on the combination of face and voiceprint.(2)Then studies the common recognition algorithm of face and voiceprint,and realizes face recognition based on the Fisherface and voiceprint recognition based on VQ respectively.Among them,the use of a modified face region histogram equalization method of face recognition in the face processing stage reduces the impact caused by uneven illumination.In addition,when the data are stored or transmitted in the data processing,the AES algorithm is introduced to encrypt the data.It can protect the biometric data from tampering and illegal attack.(3)Finally,it raises a mixed authentication design based on face and voiceprint.During the verification,it takes the fusion algorithm based on SVM to integrate many single-biometric logic output in the matching layer.The experiments prove that this method improves the efficiency of the overall identification system.Thereafter,the authentication scheme with mixed multi-level security verification form is presented.In a mixed decision-making layer,using the rule of "and" and "or" to verify.In order to verify the actual availability of the situation,the paper realizes the design to be a client of Android that uses the combination form based on third-party business library.The app set the traditional ID as the entrance.After authenticating,the user can choose a form to verify belong to the need.The various choices to login in the app include the single form based on face or voiceprint recognition and the combined form who can choose it as need.Through the analysis and test,the situation is feasible and effective.It has some advantages,such as:(1)The form of the authentication is more simple and convenient due to the result that depends on the two common biological features: face and voiceprint,compared to other traditional verification.(2)The use of an improved face region histogram equalization fusion method in the face recognition data preprocessing can improve the recognition accuracy.(3)AES encryption algorithm used in data transmission and processing can protect the data from being illegally acquired and tampering attacks,which improves the security of data and systems.(4)The proposed multi-level security mixed authentication scheme makes authentication more diverse,more innovative and more interesting.