ID R&D Report Suggests Computers Outperform People in Liveness Detection

ID R&D has published a new report that suggests that automated liveness detection is now much more reliable than the human eye. The study presented 175,000 images to human participants and an automated liveness solution, and found that people were far more likely to get it wrong when asked to determine when a real person was present in the image.

With that in mind, ID R&D argued that the results should ease some of the doubts that people tend to have about new technology, and prompt more organization to switch to automated liveness tech. Those who do so will be able to improve security and reduce administrative costs while still delivering a better end user experience.

In terms of numbers, ID R&D’s study found that its automated liveness detection tech had a zero percent error rate across all categories in the test. The dataset mixed genuine selfie images with five common forms of spoofing, including printed and digital photos, video spoofs, and 2D and 3D masks.

Humans, unfortunately, struggled with even the most basic spoof detection. They misidentified a full 30 percent of the printed photos (historically one of the easier forms of spoofing to use and spot), and they needed much more time to generate those poor results. Humans took an average of 4.8 seconds to make a decision on each image, while the computer needed less than have a second.

The human results improved when a group of 17 people were asked to vote on the authenticity of an image, though the collective was still not as accurate as the computer. Humans were also more likely to misclassify a real face as a spoof, doing so 18 percent of the time compared to one percent for the automated system. False positives are a major concern because they can add considerable friction to a user’s onboarding experience, though ID R&D believes that the performance gap still justifies the use of automatic tech.

“The results are undeniable,” said ID R&D CEO Alexey Khitrov. “Biometric technology used for identity verification has evolved in recent years to increase speed and accuracy, now significantly outperforming the human eye. Organizations can achieve tremendous efficiencies by using identity verification systems that include a biometric component.”

Mitek acquired ID R&D in 2021, citing the latter’s passive liveness technology as a motivating factor in its decision. BitcoinPoint and RelyComply are some of the organizations that have already integrated ID R&D’s IDLive Face offering into their own operations.

–

February 10, 2022 – by Eric Weiss