Research On Key Technologies Of Privacy-Preserving Biometric Online Authentication

With the rapid development of IT technology and the popularity of mobile intelligent terminals,identity authentication has gradually penetrated into people's daily lives.As an important research field of identity authentication technology,biometric identification which relies on the inherent physiological characteristics of human body such as fingerprint,face and iris to identify,has become a significant link between natural person and technology society.Compared with the traditional methods,biometric-based identity authentication is widely used in various scenarios such as public security,finance and e-commerce because of its advantages of convenience,accuracy and reliability.However,biometric authentication brings convenience to people's lives but it also leads to privacy problems.In recent years,frequently exposed database attacks,face template leaks have caused great concern in the industry,which means the protection of user privacy data is imminent.To a certain degree,whether the privacy and security of biometric data can be guaranteed during the authentication process will directly affect the future application prospects of bio-technology.Combined with the above problems,this paper studies the key technologies of privacypreserving during biometric authentication process,mainly doing the following work.First,aiming at the problem that biometric data generally lacks protection,we propose a template-based face spoofing attack.Based on the vulnerability of the face recognition system,the attack points of the system are analyzed to form the face spoofing attack ideas.After that,the genetic algorithm is used to solve the optimization problem in the attack model.The simulated attack experiment is carried out,and the optimal algorithm training parameters can obtain the ideal attack effect.Finally,through the comparative analysis of the attack results,the feasibility of system spoofing attack based on face template is proved,and the importance of bio-sensitive data protection is also revealed.Secondly,in view of the problems of computational overhead and limited application scenarios in the existing schemes,we proposes a privacy-protected online authentication scheme according to security requirements and design goals.The scheme is based on Eigenface algorithm,matrix perturbation encryption,asymmetric encryption and digital signature to ensure the security of user privacy data while taking into account the system efficiency.Subsequently,we analyzes the feasibility,security and computational complexity of our scheme.Finally,the experiment is carried out to test the privacy-preserving framework proposed in this paper.Firstly,the core algorithm performance is tested,including registration template encryption,authentication template encryption and computing system similarity score.The results show that the scheme has high computational efficiency and can be applied to multiuser and high-concurrency identity authentication scenarios.Subsequently,the authentication accuracy experiment is set.By comparing recognition performance between Eigenface algorithm and our scheme,it is verified that the application of the privacypreserving framework does not reduce the performance of the underlying matching algorithm.Finally,the experimental results show that the privacy-preserving framework proposed in this paper can provide users with a safe,efficient and accurate face online authentication service.