A new cryptographic primitive for noise tolerant template security

We introduce a new cryptographic primitive for noise tolerant template security. Our primitive consists of two algorithms. The first algorithm extracts a noise tolerant and secure template of fixed length binary data. The noise tolerance of our scheme follows from the second algorithm that determines whether two templates originate from a pair of data where the Hamming distance belongs to a priori-fixed error tolerance interval. The proposed scheme does not require any encryption infrastructure and its security does not rely on the secrecy of parameters. We discuss the security of our scheme with respect to the irreversibility and indistinguishability of templates and present concrete efficiency and security analysis for practical parameters under certain assumptions. Our techniques can potentially be used to increase security of systems and privacy of users in certain applications including biometric authentication systems and embedded security systems based on Physically Unclonable Functions. (C) 2016 Elsevier B.V. All rights reserved.