Abstract
Recently, automatic face recognition has been applied in many web and mobile applications. Developers integrate and implement face recognition as an access control into these applications. However, face recognition authentication is vulnerable to several attacks especially when an attacker presents a 2-D printed image or recorded video frames in front of the face sensor system to gain access as a legitimate user. This paper introduces a non-intrusive method to detect face spoofing attacks that utilize a single frame of sequenced frames. We propose a specialized deep convolution neural network to extract complex and high features of the input diffused frame. We tested our method on the Replay Attack dataset which consists of 1200 short videos of both real-access and spoofing attacks. An extensive experimental analysis was conducted that demonstrated better results when compared to previous static algorithms results.