Practical considerations in privacy preserving biometric face recognition algorithms

The popularity of authentication via fingerprints, iris, face or other biometric features is growing. Hence there is an increasing need to allow a wide variety of verifying parties to have access to biometric template (or reference) data. In this paper, we discuss solutions to ensure that in a biometric authentication, the verifying party cannot deduce any information from the template, the user. We are developing an experimental test bed that allows end-to-end authentication of multiple users, using techniques for privacy preserving biometric recognitio