Research On The Method Of Generating Cancelable Palmprint Features

Nowadays,biometric systems for identification or authentication are ubiquitous.These systems have many advantages,such as being difficult to lose or forgotten.In addition,biometrics provide good accuracy.However,biometric-based authentication systems face the challenge of security and privacy violations,and their theft may expose users' sensitive information.In addition,due to the invariance of the biometric,if the template is attacked,it will result in the permanent theft of the user's identity.Palmprint recognition and authentication systems also have the same security risks.Based on this,this paper mainly studies the cancelable palmprint feature generation method to effectively protect the privacy of palmprint.The following work is carried out in this paper:(1)A cancelable palmprint templates generation method based on random measurement and noise data is proposed.Firstly,Anisotropic Filter(AF)is employed to capture the orientation information of the palmprint.Then,the pseudo-randomness and non-ergodicity of the chaotic matrix are used to generate the cancelable palmprint feature.The recognition accuracy can be improved by fusing matching scores at the score-layer or the decision-layer.In this paper,the recognition performance with noise data is analyzed theoretically,and the correctness of the theory is proved by experiments.It means that we can reasonably adjusted the proportion of noise data through calculations and protect palmprint privacy on the basis of ensuring the recognition accuracy.Furthermore,our methods can still achieve very high security in the worst case of secret key theft.(2)A cancelable palmprint templates generation method based on cuckoo hash is proposed.The orthogonal features of the palmprint are extracted,and then the features are divided into blocks to perform cuckoo hash mapping to form cancelable palmprint features,forming a layer of security protection.The decimal code and gray code are used in the mapping.In addition,in order to reduce the dimension of features,the minimum hash algorithm is used to form another layer of security protection.This paper implements two-layer protection and performs a detailed security analysis of the method.The experimental results and analysis show that the scheme can effectively protect palmprint privacy.(3)A cancelable palmprint templates generation method based on matrix transformation is proposed.First,the texture features of the palmprint are extracted,then the irreversibility of the irreversible matrix is used to transform the palmprint features to form the final cancelable palmprint features.The experimental results and analysis show that the scheme can protect palmprint privacy well.