Researchers from Disney have announced a new aging/de-aging tool, Face Re-aging Network (FRAN), a neural network that can make actors look older or younger without needing weeks of complex visual effects work.
The technology will enable TV or film producers to re-age actors using an automated, cost-effective, and less time-consuming process than previous methods.
Disney presented the research in a paper called “Production-Ready Face Re-Aging for Visual Effects” and submitted it to be included at the 2022 ACM SIGGRAPH Asia conference to be held in December.
Traditionally, when special effects staff on a film or video production need to re-age the actor, they usually either use a 3D modeling and 3D scanning process or a 2D frame-by-frame digital retouching of the face using tools like Photoshop. This process can take weeks, depending on the length of the work.
In contrast, Disney’s new AI technique automates the process. Disney calls it the first production-ready, practical, and fully automatic method for re-aging faces in video images.
To build FRAN, Disney researchers randomly created thousands of samples of synthetically aged faces between ages 18-85 using StyleGAN2. With that training data in hand, FRAN learned general principles about how a person’s appearance changes with age. Now that training is complete, it can apply those aging principles to a real actor in motion, frame by frame.