Research On Privacy Protection Of Biometric Authentication Based On Feature Transformation

The progress of technology and the rapid spread of 5G networks have made accurate and fast authentication an important requirement in people’s daily lives.The traditional ways of verifying users’ identity have disadvantages such as easy to lose and forget,which causes much inconvenience in people’s life.Biometric(fingerprint)based identification method mainly relies on the physiological characteristics of individuals,which is convenient and fast without being forgotten or lost,and the technology is highly viable and practical,and has been widely used in daily life.However,some studies point out that individual biometric data is closely related to individual privacy information,and is unique and irreversible,which will cause immeasurable harm if leaked or lost.The security and privacy issues brought by biometric identification technology have aroused widespread concern.The feature templates in biometric identification systems must be protected for practical reasons.By studying and analyzing existing fingerprint feature template protection techniques,this paper investigates fingerprint authentication with privacy protection based on Diffie-Hellman key exchange and secret sharing techniques,as well as fingerprint authentication with privacy protection based on permutation transformation and Diffie-Hellman key exchange.The following is the primary work:1.In the protection of biometric templates,there is a key management problem,i.e.,whether the key is stored in the server or entrusted to the user for safekeeping,there is a risk of key leakage or loss.In order to address this issue,a privacy-preserving fingerprint authentication scheme based on Diffie-Hellman key exchange and secret sharing technology is constructed in this paper.The scheme employs the secret sharing scheme to share users’ important private information in a distributed manner around a distributed network or group,effectively reducing the storage burden on the template storage center and the user.In addition,to protect the security and privacy of the user’s fingerprint template data,the user’s original fingerprint template is stored in the template storage center in ciphertext format.When a user initiates an authentication query,a random primary key can be generated at the template storage center and the user side using the Diffie-Hellman key exchange scheme,thus achieving secondary encryption of the fingerprint template and effectively resisting man-in-the-middle attacks,replay attacks and the attacks via record multiplicity.According to the security analysis,the scheme enjoys indistinguishable against chosen-plaintext attacks and provides user anonymity.Finally,the experimental results on the public fingerprint dataset,which composes of 408 grayscale fingerprint images acquired by a Cross Match Verifier 300 sensor,demonstrate that this scheme can provide secure and accurate remote authentication services.2.Most of the existing template protection schemes focus only on the security issue of templates,while ignoring storage and efficiency issues.Based on this starting point,this paper proposes a privacy-preserving fingerprint authentication scheme based on permutation transformation and Diffie-Hellman key exchange.The scheme can reduce the system’s storage cost while ensuring the security of user templates and authentication readiness.The scheme first permutes the Finger Code feature vector corresponding to the user’s fingerprint image by using the user’s randomly selected permutation factor.Then the user’s encryption key and the selected scrambling factor are used to reduce the dimensionality of the scrambled fingerprint feature vector and generate a registered fingerprint template.When the user initiates a query,the query fingerprints undergo the same process as template fingerprints,and then the template is matched in the transformation domain by combining Diffie-Hellman key exchange.The theoretical analysis and simulation experimental results show that the scheme can provide secure,accurate and efficient authentication services.