ID R&D publishes paper to explain passive facial biometric liveness detection with a single frame image

Liveness detection with face biometrics is still a relatively new concept for most people, so ID R&D has published a white paper to explain how it can be performed with a single image and how the technology works.

The five-page “Understanding Single Frame Passive Facial Liveness” paper explains what passive facial liveness is, outlines advantages to using a single frame to confirm it, and then provides a high-level explanation for how it works. The performance of neural networks for spoof detection is then compared to humans, in so much as such a comparison is possible.

The advantages of single frame biometric liveness techniques listed by ID R&D include avoiding the friction created by requiring movements, gestures, or video to recognize liveness, the avoidance of a need for a separate interface to be integrated by developers, the imperceptibility of the process to fraudsters looking for clues on how to defeat it, and the minimization of data that needs to be uploaded from the user’s device. The company says a single 50kb image is sufficient to perform an accurate liveness check when using the recommended settings.

The latter characteristic was emphasized when South African KYC technology provider RelyComply announced in June it has selected ID R&D’s passive liveness capability.

ID R&D’s passive biometric liveness technology was confirmed compliant to the ISO/IEC 30107-3 Level 1 Presentation Attack Detection (PAD) standard in testing by iBeta announced early this year.

The company also published a white paper on the difference between active and passive liveness detection technologies earlier this year.

ID R&D executives also discussed the success of several teams from the company in a recent deepfake detection competition with Biometric Update last month.

Article Topics

biometric liveness detection  |  biometrics  |  fraud prevention  |  ID R&D  |  ISO/IEC 30107-3  |  passive facial liveness  |  spoof detection  |  white paper