Tech firm Bkav fools Apple's iPhone X Face ID

According to Bkav, all you need is to make a mask that can trick Face ID. Photo / via YouTube

Apple claims the facial recognition system on the new iPhone X is impervious to being fooled by photos, impersonators and masks, but a team of researchers claim to have beaten the technology after just a week.

Cyber security firm Bkav said a 3D-printed mask costing just US$150 ($218) to make fooled the Face ID software, which is used to unlock the iPhone X, authorise payments and log in to apps.

The researchers said it proved Face ID was "not an effective security measure", although making the mask took a detailed facial scan that would be hard for normal users to replicate.

Face ID differs from the image recognition techniques used in many other electronic devices. Photo / AP

Their demonstration has not been independently verified, and their video does not go through the entire set-up process, so there are likely to be doubts about the supposed flaw.

When the iPhone X was unveiled in September, Apple touted Face ID's security benefits, saying there was a one-in-a million chance of another person being able to unlock it.

Bkav constructed the mask using a combination of 3D printing, a silicone nose and printed images of the eyes. The video it released appears to show Face ID being fooled when a cloth covering the mask is removed, although it does not show Face ID being set up, so it cannot be confirmed that the technique works.

Face ID differs from the image recognition techniques used in many other electronic devices and which have been easily fooled merely by photos of the target. The iPhone X uses a technique called dot projection, which directs beams of infrared light at the user's face to create a 3D image, and uses artificial intelligence to "learn" the person's face.

Apple used a fingerprint sensor embedded in the home button for iPhone security for several years, but removed the home button on the iPhone X to make room for a bigger screen, leading it to develop Face ID.