Research On Hand Feature Template Protection Method Based On One Factor

With the development of information technology,biometric identification systems for identification or identity verification are widely used.These systems have the advantages of not being easily lost or forgotten,and can also provide good identification accuracy.However,biometric-based identification systems face privacy and security issues.If a biometric template is stolen,the user's private information will be exposed.In addition,due to the uniqueness and invariance of biometrics,once the biometric template is attacked,it will lead to the permanent loss of user identity.Therefore,it is necessary to protect the biometric template.Based on the above problems,this paper mainly studies two methods of protecting feature templates based on palm prints or fingerprints.First,based on palmprint features,a one-factor palmprint feature template protection method based on orthogonal index of maximum and minimum hash signature is proposed.This algorithm uses the Logistic-Tent-Sine compound chaotic model(LTSS)to generate orthogonal Gaussian random projection(GRP)matrices,then performs orthogonal index of maximum on the palmprint.The recognition accuracy,irreversibility and renewability of palmprints are improved.In addition,in order to enhance the security of the algorithm,an XOR operation is performed on the palmprint hash code obtained above and a random binary string generated by the system.Finally,the system generates a random binary string with a minimum hash signature to obtain a pseudo-identifier,and performs matching identification.Because the pseudo-identifier is generated without palmprint features,it effectively improves the security of the algorithm.Secondly,based on fingerprints,a novel one-factor fingerprint feature template protection method based on a novel minimum hash signature and a secure extended feature vector is proposed.The novel minimum hash signature is an improvement on the original minimum hash signature.It not only performs the minimum hash signature on the "1" in the fingerprint feature,but also performs the minimum hash signature on the "0".Greek codes are fused,which can extract more fingerprint information,and improve the accuracy and irreversibility of fingerprint recognition.In order to enhance the renewability and security ofthe scheme,a random binary string is added to the extended feature vector to obtain a secure extended feature vector.Finally,the above two algorithms are combined with a one factor algorithm,which enhances the security and practicability of the scheme.