Securing Biometric Data
| Citation: | Vetro, A.; Draper, S.; Rane, S.; Yedidia, J., "Securing Biometric Data", Distributed Source Coding, ISBN-13: 978-0-12-374485-2 Algorithms and Applications, January 2009 (Elsevier) |
| MERL Report: | TR2009-002 |
| MERL Contacts: | Anthony Vetro, Shantanu Rane, Jonathan Yedidia |
Robust feature extraction is combined with syndrome coding to build a secure fingerprint biometrics system.
This chapter discusses the application of distributed source coding techniques to biometric security. A Slepian-Wolf coding system is used to provide a secure means of storing biometric data that provides robust biometric authentication for genuine users and guards against attacks from imposters. A formal quantification of the trade off between security and robustness is provided as a function of the Slepian-Wolf coding rate. Prototype secure biometric designs are presented for both iris and fingerprint modalities. These designs demonstrate that it is feasible to achieve information-theoretic security while not significantly compromising authentication performance (measured in terms of false-rejection and false-acceptance rates) when compared to conventional biometric systems. The methods described in this chapter can be applied to various architectures, including secure biometric authentication for access control and biometric-based key generation for encryption.
 | Read the full technical report (PDF: 752.5 kB) |