ABSTRACT
The COVID 19 pandemic has fuelled the acceleration of the use of remote services as, for example, video conferences or digital identity verification solutions. Due to videoconferences or social medias, attackers have access to rich biometric sources and therefore make it possible to carry out high quality attacks such as videos of deepfakes, or morphing, against face recognition system. These kind of video attacks allow the attacker to fool face recognition even when these systems are secured by challenge-based liveness detection by presenting them. In order to prevent against these kind of attacks, adding an artefact detection to these systems could be a good solution. However, we will see in this paper that the development of remote digital identity verification tools on mobile application or on a computer (through a web app) opens the path to video injection attacks which bypass all of these security systems, namely a face recognition system secured with both challenge-based liveness detection and artefact detection. © by the International Institute of Informatics and Systemics.
ABSTRACT
Since the global COVID 19 pandemic, videoconference has become a daily routine for a large part of the world's population, whether for work or personal life. However, despite its many advantages, videoconference offers a significant biometric source to attackers. Indeed, we will see in this article that recording the face of a person during a videoconference can make it possible to carry out high quality attacks, in particular deepfakes and morphing, in order to attack remote facial recognition systems, secured by a challenge-based liveness detection module. © WMSCI 2021.All right reserved.