Don’t believe everything you read (or see)
We all know that we shouldn’t believe everything we read or see on TV. But if the news is showing a video of someone speaking, we can rest assured that the person really did say it, right?
Perhaps not for long. Researchers from the University of Erlangen-Nuremberg, the Max Planck Institute for Informatics and Stanford University are developing a technology that would make it incredibly easy to manipulate videos live in real time, using only a standard PC and web camera.
Manipulates videos in real time
The technology has to be seen to be believed, so check out the video here. Using software and a standard webcam, Face2Face can be used to “manipulate Youtube videos in real time,” replacing the facial expressions of the person in the target video with expressions generated by a source actor.
“No good reason for this technology other than to do nefarious s- – -“
Using a web cam, the team films the facial expressions of a source actor, which are then re-rendered to simultaneously replace the original facial expressions of the person in the video. Similar technologies have been used to this effect before, but have rarely been convincing (see Beyoncé’s “Grown Woman” video), and the technique has never been applied in real time.
The improvement comes from a more accurate re-rending of the interior of the actors mouth, such that the movements of the source actor’s mouth are matched with the mouth of the target video.
The project seems to arouse both excitement and suspicion. Youtube users’ responses to the project range from “wish I could go back to grad school to play with technology” to “there is no good reason for this technology other than to do nefarious s—.”
Purely research-focused (for now)
The applications of such a technology are, undeniably, a bit frightening to think about. The research team seems aware of the potential sketchiness factor, explaining defensively that their project is “purely research-focused.” They claim that they have no sinister purpose, but are only hoping to “demonstrate the capabilities of modern computer vision and graphics technology….in an approachable and fun way.”