Privacy Protection in High Security Biometrics Applications

Abstract

While biometrics is useful for secure and accurate identification of a person, it also has serious privacy implications in handling large databases of biometrics. Standard encryption techniques have limited use for handling biometrics templates and signals as intra-person variations can’t be handled effectively using information security techniques. As a way to enhance privacy and security of biometrics databases, we present a pattern recognition-based model to analyze the threats to a biometrics-based authentication system and also a novel solution to enhances privacy. Cancelable biometrics is an emerging concept, where transformations that hide the biometrics signatures (fingerprints, faces and iris) are designed so that the identity can be established with added security. Other methods have been proposed for enhancing privacy in biometrics. In this paper, we will describe our threat model for biometrics recognition and recent advances proposed for privacy enhancements for fingerprint and iris biometrics.