Vulnerabilities of Unattended Face Verification Systems to Facial Components-based Presentation Attacks: An Empirical Study

As face presentation attacks (PAs) are realistic threats for unattended verification systems, attack detection (PAD) has been intensively investigated in past years, and the recent advances PAD have significantly reduced success rate of such attacks. In this article, an empirical study on a novel effective impostor PA is made. proposed PA, facial artifact created by using most vulnerable components, which optimally selected based vulnerability analysis different components to PAs. An attacker can launch presenting his or her own real face. With collected database containing various types artifacts instruments (PAIs), experimental results show that poses more serious threat systems compared with print, replay, mask Moreover, generalization ability regard commercial also evaluating unknown real-world systems. It provides new paradigm