Learning Deep Feature Representation for Face Spoofing

Biometrics is an emerging research area due to its easiness in identification of the person. Face Spoofing is the challenging task in face recognition systems because the human can easily trickster the system by presenting the video or photograph of the person. Many approaches are providing good results in face spoofing, but still it is challenging in intra and cross database validation. Deep learning algorithms have shown significant results in the intra and cross database. This paper used deep learning for extracting the inclusive and favorable features of the person from the face. The extracted features are used for classifying the face image as a real face or genuine face. The performance of the method is evaluated through statistical measures. The experiments were carried out NUAA and CASIA database. The method attained most promising results than other face spoofing methods.