Research On The One-factor Cancelable Biometrics Template Protection Algorithm

Biometrics,as one of the new technologies of identity verification,has been widely used in various applications such as economy,personal terminals,national defense security and e-government.Due to the unique characteristic of biometrics,once the user's biometrics are stolen,the leakage of biometrics is permanent.Therefore,the study on the template protection algorithm of biological characteristics has crucial significance.The one-factor cancelable biometric template protection is a new template protection approach,which only needs the biometric factor as the input.It solves the problem that the external factors in the two-factor cancelable biometric template protection may be stolen,lost or forgotten.Yet,it is difficult for the scheme to meet the requirements of irreversibility,non-linkability,performance maintenance and security simultaneously.In view of the above challenges,this thesis has done the following two aspects of research work based on one-factor cancelable biometric template protection of the fingerprint scene and the finger vein scene.On the one hand,the current the one-factor cancelable fingerprint template protection algorithm with better recognition performance cannot guarantee security,but the recognition performance of the secure algorithm is poor.In order to improve the performance and security of the one-factor cancelable fingerprint template,a One-factor Window Sliding and Extracting Hashing(WSE)algorithm using fingerprint is proposed.The algorithm takes fingerprint features as the only input and improves the method of feature extraction.Firstly,using extended feature vectors,binary seeds are generated by predefined sliding window and hash function randomization;then different auxiliary data are replaced to generate revocable templates;finally,the auxiliary data is decoded by querying the biological feature vectors to improve the irreversibility and security.The experimental results in fingerprint database FVC2002 and FVC2004 show that this method not only meets the four design criteria of cancelable biometric template,but also defends three kinds of security attacks.On the other hand,this thesis proposes One-factor Blocking and Dimensionality Reduction Hashing(BDR)algorithm based on finger vein.The algorithm takes the characteristics of the finger vein as the only input,and the size of the database is n ?m according to the characteristics of the finger vein,Firstly,the features of finger vein are divided into blocks,and each block is transformed and extracted respectively,then generating 1-dimensional binary vector from the extracted feature vectors by concatenation function,to achieve dimensionality reduction;Then the feature vectors after dimensionality reduction are transformed and improved hash operations are used to generate the protected finger vein template;finally,the revocable finger vein template is generated together with the embedded auxiliary data.The experimental results of UTFVP in finger vein database show that the method meets the criteria of cancelable biometric template,and the security is analyzed.This thesis proposes an improved one-factor cancelable biometric template protection algorithm for fingerprint and finger vein.The experimental results on their respective feature databases show that the algorithms satisfy the requirements of cancelable biometric template protection,and strong security.At the same time,it has important practical significance to reduce storage space and improve the efficiency of computing.