Using a challenge to improve face spoofing detection

Fokkema, J. (2016) Using a challenge to improve face spoofing detection.

[img]
Preview
PDF
2MB
Abstract:With biometric face recognition increasing, there comes a natural increase in interesting attack scenarios for people with bad intentions. One way to trick biometric recognition methods is to spoof the system by showing an artificial reproduction of somebody else his face. This can be done by e.g. showing a printed picture, a video or by wearing a mask. Multiple methods are developed to detect this kind of spoofing, but these are still lacking performance in terms of detection rates. Also, many of these methods are not applicable to mobile devices, due to specific constraints that exists on these devices. In this paper we extend an existing spoofing detection method with usability on mobile devices in mind. The existing method uses 3D properties of the face to distinguish between a real face and 2D representations of a face. The extension on this method is a challenge, which requires the user to rotate its head in a specific pattern. By adding this challenge, the necessary rotation for the 3D check is enforced and furthermore, theoretically, more types of spoofing can be detected. Tests shows that users can successfully perform the challenge and that it is capable of reliably detecting spoofing attacks.
Item Type:Essay (Master)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science MSc (60300)
Link to this item:http://purl.utwente.nl/essays/71438
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page