Abbreviation view of face presentation attack techniques

Abstract

The popularity of biometric technology is rising and has become a part of our everyday lives. The biometric system aims to recognize human characteristics instantly. In an operating framework context, between the biometric solutions deployed, the use of facial biometrics plays a significant role. In the biometric society, attacks have developed great attention (also known as spoofing attacks). Although the tremendous improvement in the recognition of human faces, existing systems remain prone to presentation attacks, subverting the face recognition systems by showing a fake face item. The biometric system sensor is the easiest to attack where it is the most exposed system component. The attacks of Biometric sensors are named Presentation Attacks (PA). Several methods have been developed to automatically detect different presentation attacks, primarily for video replay and 2D photo printing attacks. These have shown a high-security threat for systems of face recognition. However, many anti-spoofing techniques for presentation attack detection (PAD) have been suggested that can identify and reduce such targeted strikes automatically. This survey is carried to provide an abbreviation review of the latest work that has been accomplished on the detection of face presentation attack. It discusses the various types of face presentation attacks, threat assessments and metrics of evaluation of output effects, and public database availability to benchmark new PAD algorithms, as well as an outline of the relevant research in this field. Besides, in this emerging area of biometrics, also explore the open issues and potential work that needs to be tackled.