Biometric layering: template security and privacy through multi-biometric template fusion

摘要: As biometric applications are gaining popularity, there is increased concern over the loss of privacy and potential misuse data held in central repositories. Biometric template protection mechanisms suggested recent years aim to address these issues by securing a or other structure such that it suitable for authentication purposes, while being protected against unauthorized access crosslinking attacks. We propose framework enhancing security, layering multiple modalities construct multi-biometric difficult extract separate individual layers. Thus, uses subject's own conceal her data, also enjoys performance benefits because use modalities. The resulting cancelable if system implemented with biometrics as voice. present two different realizations this idea: one combining fingerprints another fingerprint spoken passphrase. In either case, both samples required successful authentication, leading addition gains. proposed evaluated using FVC 2000-2002 NIST databases, TUBITAK MTRD speaker database. Results show only small degradation EER compared state-of-the-art ngerprint verification high identification rates, cross-link rates low even very databases.