Font Size: a A A

The Research On The Encryption Of Finger Vein

Posted on:2021-12-11Degree:MasterType:Thesis
Country:ChinaCandidate:J ZhangFull Text:PDF
GTID:2518306308467954Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
In the field of identity authentication,biometric technology has become an ideal solution for complex and high security demand scenarios.Finger vein recognition,as the second-generation biometric technology,has both the advantages of high security,high speed,living body recognition,convenience,and high accuracy,so it has been rapidly promoted.However,the biometrics such as finger veins are vulnerable and difficult to cancel.In order to reduce the risk of personal privacy leakage,an effective solution is to encrypt the original biometric template before storing.Traditional image encryption algorithms are difficult to resist malicious key sharing and non-repudiation attacks.Therefore,biometric transformation technology and bio-cryptosystem came into being.The bio-cryptosystem is a powerful combination between biometrics and cryptography where biometrics eliminate the need to remember passwords,and cryptography technology provides a higher security level for biometrics.The work of this paper first analyzes the research background,current situation and difficult problems of biometric encryption.Secondly,the biometric template protection technology is reviewed.Finally,three robust finger vein encryption schemes are proposed around feature transformation and bio-cryptosystem,and a finger vein feature extraction scheme based on scale-invariant feature transform is also proposed.The main research contents and work results of this article are as follows.(1)Finger vein encryption scheme based on Tree-BioHashing.It has been verified by experiments that the recognition performance of the finger vein encryption scheme based on the original BioHashing is severely degraded when the token is leaked.This paper proposes a Tree-BioHashing algorithm that combining original quantization and binary tree quantization.In the authentication stage,an adaptive distance measurement scheme is designed to solve the ambiguity of biometric.The experiment found that when the token is secure,the EER of this scheme is 0;when the token is leaked,the recognition performance is significantly improved compared to the original BioHashing.(2)Two-factor and cancellable finger vein encryption scheme based on Fuzzy Vault.Firstly,the experiment analyzes that the existing Fuzzy Vault is vulnerable to correlation attack and short-key attack.This paper proposes two solutions to the correlation attack,which are two-finger veins construction and one-way transformation for genuine points.For short-key attack,the user token is hashed to generate a long key with sufficient strength in advance,and the key is shuffled.The experiments prove that the scheme had good recognition performance and robustness.(3)Finger vein feature extraction scheme based on Grid-SIFT.First,the experiment verifies that the finger vein feature based on SIFT has robustness to scale,brightness,rotation transformation.However,the keypoints of SIFT have the defects of high descriptor dimension,inefficient matching and inapplicability to certain bio-cryptosystems.Therefore,this paper designs a point-to-string conversion framework based on SIFT keypoints.The main work is to meet the real-time requirements of rapid matching and the extensive requirements for bio-cryptosystems.(4)Finger vein encryption scheme based on Grid-SIFT FE.Some bio-cryptosystems rely on error correction codes to handle intra-class differences in biometrics.However,reducing the inherent uncertainty is more direct and effective than relying on error correction codes to correct the ambiguity of biometrics.Therefore,this solution uses Grid-SIFT to extract finger vein features with discriminating ability.Then Fuzzy Extractor is used to encrypt finger veins and generate keys.Finally,the feasibility of the scheme is proved from performance experiments and safety analysis.The applications of finger vein recognition are accelerating,but the research on finger vein encryption is not yet mature.The research in this paper can provide more feasible research ideas for finger vein encryption based on biometric transformation and bio-cryptosystem.
Keywords/Search Tags:Finger Vein Recognition, Finger Vein Encryption, Biometric Transformation, Bio-cryptosystem
PDF Full Text Request
Related items