Anti-spoofing: Fingerprint Databases

Anti-spoofing may be defined as the pattern recognition problem of automatically differentiating between real and fake biometric samples produced with a synthetically manufactured artefact (e.g., latex or silicone finger). As in any other machine learning problem the availability of data is a critical factor to be able to successfully address this challenging task. Furthermore, these data should be public, so that the performance of different protection methods may be compared in a fully fair manner. The present entry describes general concepts regarding spoofing dataset acquisition and particularizes them to the field of fingerprint recognition. It also gives a summary of the most important features of the current publicly available fingerprint spoofing databases.JRC.G.6-Digital Citizen Securit