SIGGRAPH is coming, when all the amazeballs graphics research drops, and the previews are terrifying and astonishing by turns (sometimes both!).
A team led by Stanford's Michael Zollhöfer has posted a preview of their research on "Deep Video" — using deep learning algorithms to edit existing videos so that the words and gestures of one person can be mapped onto another person.
We've seen this before (remember last year's doctored Obama clips?), but Zollhöfer's team has done some truly next-level shit here, with a suite of spookily good effects for faking videos that target the deficiencies in earlier methods that might have helped sleuths, journalists and the public tell real from fake.
We present a novel approach that enables photo-realistic re-animation of portrait videos using only an input video. In contrast to existing approaches that are restricted to manipulations of facial expressions only, we are the first to transfer the full 3D head position, head rotation, face expression, eye gaze, and eye blinking from a source actor to a portrait video of a target actor. The core of our approach is a generative neural network with a novel space-time architecture. The network takes as input synthetic renderings of a parametric face model, based on which it predicts photo-realistic video frames for a given target actor. The realism in this rendering-to-video transfer is achieved by careful adversarial training, and as a result, we can create modified target videos that mimic the behavior of the synthetically-created input. In order to enable source-to-target video re-animation, we render a synthetic target video with the reconstructed head animation parameters from a source video, and feed it into the trained network — thus taking full control of the target. With the ability to freely recombine source and target parameters, we are able to demonstrate a large variety of video rewrite applications without explicitly modeling hair, body or background. For instance, we can reenact the full head using interactive user-controlled editing, and realize high-fidelity visual dubbing. To demonstrate the high quality of our output, we conduct an extensive series of experiments and evaluations, where for instance a user study shows that our video edits are hard to detect.
Deep Video Portraits [H. Kim, P. Garrido, A. Tewari, W. Xu, J. Thies, M. Nießner, P. Perez, C. Richardt, M. Zollhöfer and C. Theobalt/Siggraph 2018]
(via Four Short Links)