A New design approach of Fuzzy Hashing schemes for fingerprints

Abstract

Fuzzy hashing is a technique commonly used to compare two diff erent data les which share identical sets of bits in same order, by determining the level of similarity between them. This is done by generating same hash output for the similar les (similar in terms of bits). We provide a new design approach of Fuzzy Hashing schemes, for protecting fingerprint templates of individuals from being misused by attackers when stored on servers or any Cryptographic module by introducing a binary representation for templates. The representation scheme outputs a bit string from the unique features present on human fingertip and is non-invertible, invariant to change in position of fingerprint, and captures most of the discriminatory information which can de fine a fi ngerprint. As a particular individual has unique set of features known as minutiae, the bit string generated is unique for every individual. The user is authenticated through simple hamming weight calculation between two bit strings representing fingerprint templates. We then mapped this bit string obtained from design approach to a fuzzy extractor which is used to generate strong, secure Cryptographic key that is used for encrypting personal data.