Want to keep the TNW Conference vibe going?? Tickets for TNW2022 are available now >>

The heart of tech

This article was published on August 1, 2017

This video shows the eerie perfection of real-time facial expression copying

This video shows the eerie perfection of real-time facial expression copying
Tristan Greene
Story by

Tristan Greene

Editor, Neural by TNW

Tristan covers human-centric artificial intelligence advances, politics, queer stuff, cannabis, and gaming. Pronouns: He/him Tristan covers human-centric artificial intelligence advances, politics, queer stuff, cannabis, and gaming. Pronouns: He/him

A group of developers have managed to create a VR technology that let’s you put your expression on someone else’s face. It’s called Face VR and it’s a method for gaze-aware real-time facial reenactment in virtual reality — basically a way to imitate live facial expression in VR.

The team behind Face VR is comprised of researchers from several different German universities. The new tech, which builds on work the team debuted last year, will be shown off at the SIGGRAPH event this week.

This technology relies on custom algorithms to process tiny movements, including eyeball and eyelid movements. What makes this most interesting, aside from its amazing ability to track the tiniest details of your expression in real-time, is that it can render those expressions on a face other than your own.

Credit: Face VR

The researchers envision the final product as an upgrade to ‘avatars’ in VR environments. The team views Face VR as a way for people to achieve greater immersion for business meetings and social interactions using the VR platform. They also see the entertainment value behind the technology:

In addition, we plan to keep track of the best and funniest reenactment results, and share them with the world via a social media platform

The creators of this tech might look at it as a tool for immersive VR meetings, but it’s just spooky enough to remind us that we are sprinting ever-closer to being incapable of determining whether we’re talking to a person, a computer, or both.