Multilevel fusing paired visible light and near-infrared spectral images for face anti-spoofing

摘要: Abstract Face anti-spoofing has become a vital element for guaranteeing the security of face recognition systems. Previous approaches generally exploit cues in visible light images or near-infrared individually. Few studies pay attention to fusing and anti-spoofing. However, strengths weaknesses can be complementary. In this study, we introduce new dataset named as CIGIT-PPM, which includes paired with spoofing medium, distance, pose, expression session variations both print 3D mask attacks. Further, propose novel spectral CNN based approach anti-spoofing, combining discriminating ability images. Specifically, an end-to-end is employed learn representation from do classification at multilevel, then weight averaged strategy utilized integrate probability different levels, gives final result that input are captured live spoof face. Extensive experiments show our method achieves state-of-the-art results on self-collected CIGIT-PPM public msspoof.