Turn your smartphone to portrait mode

Turn your smartphone to portrait mode
to view our website correctly, thank you!

October 26, 2022

Remove the hassle, not the mask

As festive holiday travel picks up, people are increasingly concerned about safety protocols, given the ongoing presence of Covid variants, particularly in densely populated spaces with high traffic volumes and the requirement to touch shared infrastructure. Businesses, large and small, are continuing to reevaluate their safety measures and responding by implementing more contactless experiences. However, in areas where the public is still recommended to wear a mask, some facial recognition technology is not sophisticated enough to authenticate a face with the mask on. Think of access gates that allow entry into buildings and other secure areas – some technologies still require people to lower their face masks. This limitation in many facial recognition solutions inhibits the customer experience, decreases public safety, and slows down queues.

Authentication while wearing a mask

Most facial recognition technologies have a low similarity score when an individual wears a mask simply because their full face is not visible. To overcome this, VisionLabs technology provides industry-leading liveness checks securely detecting that the person being scanned is actually present, even while wearing a mask.

What goes on behind the lens?

To truly understand how VisionLabs maximizes safety and customer experience, it is essential to look at what happens behind the lens. VisionLabs has moved away from outdated active liveness checks requiring customers to respond to a task to confirm validation. Instead, it has developed state-of-the-art one-shot liveness that authenticates a person without relying on movements such as blinking or swallowing.

A depth liveness check is also incorporated into VisionLabs’ facial recognition technology. The technology assesses how far away images being scanned are – down to the millimeter – meaning it ‘understands’ if the image is flat like a video or photo, rather than a real person.

VisionLabs’ software is consistently top-rated in the NIST FRVT* benchmark, with an accuracy of 99.9% for both one-to-one (1:1) and one-to-many (1:N). Thus, when a person is wearing a mask, the level of identity accuracy is best-in-class and still superior to most competitors.

These combined elements ensure that VisionLabs responds to the current landscape’s need for safer protocols. Businesses implementing these highly accurate technologies can offer their customers and employees unparalleled, seamless digital experiences. For businesses, the benefits are numerous: securer operation lends itself to less fraud, efficient processes and fewer costs, while greater accuracy means better customer and employee experience.

*It does not imply any recommendation or endorsement by the National Institute of Standards and Technology.