Face presentation attack detection with passive imaging at 250 GHz

Systems for detecting face presentation attacks are facing great challenges since new 3D facial masks are used. Passive terahertz imaging offer specific physical properties that may improve presentation attack detection capabilities. The aim of this paper is to present outcomes of study on using novel 3D facial masks for face presentation attack detection in passive imaging at 250 GHz. A thorough analysis of presentation attacks for facial recognition systems using custom 3D-printed and custom flexible 3D-latex masks is provided together with spectral characterization of various items. A set of experiments with various spoofing items is described and discussed. Finally, a presentation attack detection method with passive imaging at 250 GHz is presented.