Impostor detection using facial stereoscopic images

Biometric systems are prone to impostor attacks, which in case of high security applications could be critical. These attacks range from simple attacks consisting of stealing someone´s password or complex as cryptographic attacks on the transmission channel. In this study we have focused on combating replay attacks, in which an impostor uses a prerecorded image of the client. The system captures stereoscopic images of the face and then decides if the images belong to a real face or a poster from depth calculated using triangulation.