Privacy Preserving Fingerprint Identification

Abstract

Privacy preserving technologies are used to create applications for computing on sensitive data without compromising on the secrecy of said data. In this thesis, secret sharing based multi-party computation is used to identify a fingerprint sample amongst a database of templates while preserving the secrecy of the sample and the templates. The FingerCode representation of fingerprints is used. Privacy preserving fingerprint identification mitigates some of the privacy and security risks in biometric identification systems. The secret sharing based fingerprint identification application developed in this thesis is more performant than a previous homomorphic encryption based one. Methodology for identifying fingerprints and programming privacy preserving applications using multi-party computation is given. Fingerprint-based identification systems are vital tools for border control and law enforcement. Privacy preserving fingerprint identification could be used to prevent leakage and abuse of fingerprint data.