This paper addresses the problem of face presentation attack detection using\ndifferent image modalities. In particular, the usage of short wave infrared\n(SWIR) imaging is considered. Face presentation attack detection is performed\nusing recent models based on Convolutional Neural Networks using only carefully\nselected SWIR image differences as input. Conducted experiments show superior\nperformance over similar models acting on either color images or on a\ncombination of different modalities (visible, NIR, thermal and depth), as well\nas on a SVM-based classifier acting on SWIR image differences. Experiments have\nbeen carried on a new public and freely available database, containing a wide\nvariety of attacks. Video sequences have been recorded thanks to several\nsensors resulting in 14 different streams in the visible, NIR, SWIR and thermal\nspectra, as well as depth data. The best proposed approach is able to almost\nperfectly detect all impersonation attacks while ensuring low bonafide\nclassification errors. On the other hand, obtained results show that\nobfuscation attacks are more difficult to detect. We hope that the proposed\ndatabase will foster research on this challenging problem. Finally, all the\ncode and instructions to reproduce presented experiments is made available to\nthe research community.\n