Face liveness detection under bad illumination conditions

Spoofing face recognition systems with photos or videos of someone else is not difficult. Sometimes, all one needs is to display a picture on a laptop monitor or a printed photograph to the biometric system. In order to detect this kind of spoofs, in this paper we present a solution that works either with printed or LCD displayed photographs, even under bad illumination conditions without extra-devices or user involvement. Tests conducted on large databases show good improvements of classification accuracy as well as true positive and false positive rates compared to the state-of-the-art.