How to use bit-string representation of the face in a multi-factor authentication system

Abstract

Binary information extracted from biometrics is variable when it is generated from photographic images. If the same bit-string representation can be produced repeatedly, these bit-strings are useful for applications that require binary data (e.g., biometric cryptosystems). In this paper, multi-factor authentication using bit-string representation is proposed. For the protection of biometric information, random bits were used as the representation, and the length of the bit-string representation was decided according to the similarity level. The paper explains what “similarity enhancement” is and how to increase the similarity between images of the same person with weights added after verifying each factor in a multi-factor authentication system. Also, it discusses how to use this method in the real world and provide secret sharing using this multi-factor system. In an experiment with the Yale Face Database A and B+, it was shown that similarity enhancement could be achieved, and the same bit-string representation emerged after passing the multi-factor authentication system.