Spoof attack by replicating biometric traits represents a real threat to an automatic biometric verification/ authentication system. This is because the system, originally designed to distinguish between genuine users from impostors, simply cannot distinguish between a replicated biometric sample (replica) from a live sample. An effective solution is to obtain some measures that can indicate whether or not a biometric trait has been tempered with, e.g., liveness detection measures. These measures are referred to as evidence of spoofing or anti-spoofing measures. In order to make the final accept/rejection decision, a straightforward solution to define two thresholds: one for the anti-spoofing measure, and another for the verification score. We compared two variants of a method that relies on applying two thresholds – one to the verification (matching) score and another to the anti-spoofing measure. Our experiments carried out using a signature database as well as by simulation show that both the brute-force and its probabilistic variant turn out to be optimal under different operating conditions.
We address the problem of score level fusion of intramodal and multimodal experts in the context of biometric
identity verification. We investigate the merits of confidence based weighting of component experts. In contrast
to the conventional approach where confidence values are derived from scores, we use instead raw measures of
biometric data quality to control the influence of each expert on the final fused score. We show that quality based
fusion gives better performance than quality free fusion. The use of quality weighted scores as features in the
definition of the fusion functions leads to further improvements. We demonstrate that the achievable performance
gain is also affected by the choice of fusion architecture. The evaluation of the proposed methodology involves
6 face and one speech verification experts. It is carried out on the XM2VTS data base.