Facial Manipulation Integration

From Quartz Media

A German team of Computer Scientists reveal their public software. Using 2 video streams or one live feed and one target video. 3d modeling and no 3d cameras, you get real time alteration. To manipulate any face in feeds target video clip, altering the subject's expressions and words by mapping another video feed's movements and superimposing those motions onto the target video.

https://www.engadget.com/2018-06-05-ai-transfer-facial-movements-from-one-video-to-another.html

paper on similar tech:
https://openaccess.thecvf.com/content_CVPRW_2019/papers/Media%20Forensics/Sabir_Recurrent_Convolutional_Strategies_for_Face_Manipulation_Detection_in_Videos_CVPRW_2019_paper.pdf

Cornell paper on the wider trust implications and dangers of this:
https://arxiv.org/abs/2112.10936

Yeah, for foreign language film dubbing... yeah ... Italians mastered that 40 years ago with simple timing. Watch Luchino Visconti’s The Leopard, Burt Lancaster look like he's speaking Italian even though he's acting in his native English.

And this is just what's public... think what they've had since before even TV was released. This tech ain't new, just new for us...