Disney Research and ETH Zurich introduced an algorithm for automatic face-swapping in photos and videos, which has a high enough resolution for filmmaking. It's the first deepfake method with a megapixel resolution, meaning it's convincing enough to alter faces for TV and movies, according to the researchers.
- The goal is to use the neural-network-based method to substitute actors' faces, such as in the de-aging process or when an actor has died.
- Disney’s model can create deepfake videos with a 1024 x 1024 resolution, much higher than open-source model DeepFakeLab's 256 x 256 pixels.
- It's less time-consuming and expensive than Disney's current method for face-swapping, traditional VFX, which was used to re-create Carrie Fisher in Rogue One.
- TechCrunch's Darrell Etherington described the results as "a lot less nightmare fuel" than other deepfake attempts. He still noted that the subjects in the sample video were primarily white.
- The paper, “High-Resolution Neural Face Swapping for Visual Effects,” will be presented at this week's virtual 2020 Eurographics Symposium on Rendering.