Personal identity verification by serial fusion of fingerprint and face matchers

摘要: The use of personal identity verification systems with multi-modal biometrics has been proposed in order to increase the performance and robustness against environmental variations fraudulent attacks. Usually fusion is performed parallel at score-level by combining individual matching scores. This strategy exhibits some drawbacks: (i) all available are necessary perform fusion, thus time depends on slowest system; (ii) users could be easily recognizable using a certain biometric instead another one (iii) system invasiveness increases. A characterized serial combination multiple can good trade-off between time, acceptability. However, these have poorly investigated, no support for designing processing chain given so far. In this paper, we propose novel scheme simple mathematical model able predict two serially combined matchers as function selected chain. Our helps designer finding allowing trade-off, particular, time. Experiments carried out well-known benchmark data sets made up face fingerprint images usefulness methodology compare it standard fusion.