Face liveness detection using analysis of Fourier spectra based on hair

The existing face liveness detection algorithms have satisfying performance on spoofing photos by detecting the biological motion, including eye blinking, head rotation and mouth movement. However, the biological motions can be simulated by using video, which is called a video playback attack. This paper proposes a countermeasure to a video playback attack by analyzing the detail of hair using the Fourier spectra. We apply a flashlight to enhance the difference between the real person and the video playback attack. Experimental results suggest that the proposed method has an encouraging performance on face liveness detection.