Liveness detection via oculomotor plant characteristics: Attack of mechanical replicas

A novel approach that performs liveness detection for biometric modalities that use eye movement signal for person identification is proposed and evaluated. Liveness detection is done via estimation and analysis of the internal non-visible anatomical structure of the human eye termed Oculomotor Plant Characteristics (OPC). At this stage of its development the OPC approach targets prevention of spoof attacks that are generated by the accurate mechanical replicas of the human eye. We generalize and test two classes of such eye replicas via their mathematical representations. Specifically, we investigate following classes of replicas: a) those that are built using default OPC values specified by the research literature, and b) those that are built from the OPC specific to an individual. The results that involved processing live data from 32 individuals over four recording sessions and their eye replicas indicate relatively high theoretical resistance of the OPC liveness detection method to the mechanical attack that impersonates an authentic user.