Using Visual Rhythms for Detecting Video-Based Facial Spoof Attacks
The identification of video source is very important for video validation evidence, tracking down video piracy crimes and regulating individual video sources. User authentication is an important step to protect information and in this context, face biometrics has more advantage. Face biometrics are natural, intuitive, easy to use and less human invasive. Unfortunately, recent work has face biometrics vulnerable to spoofing attacks using cheap low-tech equipment. We have introduced a method for face spoofing detection using spatiotemporal (dynamic texture) extensions of highly popular local binary pattern operator. With wide deployment, face recognition systems has been used in applications from border control to mobile device unlocking and laptop device unlocking. The combat of video spoofing attacks requires increased attention. We address the problem of video spoofing detection against replay attacks by using the aliasing analysis in spoofed face videos. We analyse the texture pattern aliasing that commonly appears during recapture of video or photo replays on screen in different channels (R, G, B and grayscale) and regions. Multi-scale LBP and SIFT features determines the texture patterns characteristics which differentiate a replayed spoof face from a live video (face present). We have introduced effective approach in face spoof detection both cross-database, and intra-database testing scenarios(video) and shows better comparison since we compare the edge pixel values and depth of pixel values of the authenticated person with the image stored in the database.