Nomad Biometric Authentication
For efficient practical utilisation, biometric authentication systems should be user friendly, fast and reliable. In practice however, many prove to be intrusive and cumbersome with varying levels of reliability. Iris recognition, for example, is potentially the highest performing solution, but it typically relies on high resolution, near infrared, images captured under very controlled conditions. Moreover, these systems require the biometric data of individuals to be stored on servers, which may conflict with the protection policies for personal data. One solution is to use multimodal nomadic authentication. The advantages are numerous: improved usability, localized storage of private data, low-cost, personalized authentication.
The objective of the NOBA project is to demonstrate the feasibility of an unconstrained mobile nomadic system that creates strong authentication by acquiring and fusing biometric data. The primary biometric modalities adopted are iris (for close range, high resolution acquisition); face (for mid-range, mid resolution acquisition); and gait (for long range, low resolution acquisition). Emphasis is given to quality assessment and adaptive techniques to generate robust and reliable authentication in unconstrained scenarios where many factors may impede the acquisition of ideal samples.
Further information is on the joint project website.