Research And Design Of Encryption Scheme Of Facial Feature Template

With the rapid development of information technology and the advent of the global Internet era,user identity authentication methods based on biometric identification technology have been widely used.Compared with the traditional password,chip card and other methods,the biological characteristics have the advantages of memory-less and more safety and reliability.In many biometric recognition systems,face recognition has the advantages of not relying on special hardware support,consistent with human recognition habits,and so on.With more and more APPs and hardware devices adopting the identity authentication method based on facial feature recognition technology,the security of the face recognition system has attracted people's attention.Once the facial feature template data is leaked,we cannot report and unbind it like a bank card,and the losses caused will be difficult to estimate.In any face-based identity authentication system,the importance of protecting feature templates is self-evident.However,most of the existing biometric template encryption schemes are based on iris and fingerprints with relatively stable features,which is not effective when applied to face recognition systems with high ambiguity.This subject aims to study a robust,practical,and biometric encryption scheme that can be used for high-ambiguity facial features.In the paper,we first introduced the research background,purpose and significance of facial feature encryption technology,outlined the development process of biometric encryption technology from 1994 to the present,analyzed the current research status,and carried out the classic biometric encryption technology.The design of the biometric encryption scheme must be based on a specific biometric extraction algorithm.In this article,we introduced the principles of the PCA,DCNN,and LBP facial feature extraction algorithms,and implemented the LBP facial feature extraction and DCNN facial key point extraction.After that,we introduced the most classic fuzzy vault algorithm in detail,and explained the error correction code detection technology,hash point generation technology and polynomial reconstruction technology in detail.At the same time,we improved the fast generation method of hash points,our improved solution not only requires a sufficient distance between the hash point and the real point,but also needs to ensure that the distance between each hash point also meets the fuzzy threshold requirement.The simulation experimental results show that the algorithm is fast and effective.Finally,we propose a fuzzy encryption scheme based on 83 face key point features,which uses an improved hash point generation algorithm.We collected real face data in the experiment and used them to test the fuzzy encryption scheme.The experimental results show that the scheme improves the fuzzy tolerance and recognition efficiency,and the system's false recognition rate and rejection rate are reduced to 13.28% and 10.89 %.In the process of analyzing the security of the program,we explained the inherent security vulnerabilities of the fuzzy vault program,and envisaged an improved idea for this.