Abstract

Face Recognition is widely used in security systems, such as surveillance, gate control systems, and guard robots, due to their user friendliness and convenience compared to other biometric approaches. Secure face recognition systems require advanced technology for face liveness detection, which can identify whether a face belongs to a real client or a portrait. However, with the development of display devices and technology, the tools and skills for carrying out spoofing attacks with images and videos have gradually evolved. In this paper, we compare real faces with high-definition facial videos from LED display devices, and present the changes in face recognition performance according to lighting direction.