Tongxin xuebao (May 2020)
Efficient and privacy-preserving online face authentication scheme
Abstract
In traditional face authentication system,the trait template and authentication request were generally matched over plaintext,which may lead to the leakage of users’ sensitive data.In order to address the above-mentioned problem,based on matrix encryption,an efficient and privacy-preserving online face authentication scheme was proposed.Specifically,the users’ face trait template for register and the authentication request were encrypted before being sent to the online authentication server,and the similarity computation between the encrypted face trait template and authentication request was computed by the online authentication server over ciphertexts,which guaranteed the security of users’ sensitive data without affecting the accuracy of face authentication.Security analysis shows that the proposed scheme can achieve multiple security levels according to different security parameters.Moreover,performance evaluation shows that the proposed scheme has low computation cost and communication overhead.Experiments results demonstrate the high efficiency of the proposed scheme,which can be implemented in the real environment effectively.
