Disney’s new neural network can easily change a player’s age

Enlarge / An example of Disney’s FRAN age-changing AI showing the original image on the left and re-aged rows of older (top, 65 years old) and younger (bottom, 18 years old) samples of the same person.

Disney researchers have created a new neural network that could change the visual age of actors in TV or film, Gizmodo reports. The technology will enable TV or film producers to make actors appear older or younger using an automated process that will be less costly and time consuming than previous methods.

Traditionally, when special effects personnel in a video or film production need to make an actor appear older or younger (a technique Disney calls “re-aging”), they usually use either a 3D scan and 3D modeling process or a 2D frame. -Digital retouching of the player’s face using tools similar to Photoshop. This process can take weeks or longer, depending on the length of the job.

In contrast, Disney’s new artificial intelligence technique called Face Re-aging Network (FRAN) automates the process. Disney calls it “the first practical, fully automated, production-ready method for re-aging faces in video footage.”

A research promotional video for Disney’s re-aging technology.

To create the FRAN, Disney researchers used StyleGAN2 to randomly generate thousands of samples of synthetically aged faces aged 18 to 85 years. With this training data in hand, FRAN learned the general principles of how a person’s appearance changes with age. Now that the training is complete, he can apply these aging principles to a real player in action frame-by-frame.

“Our network was supervisedly trained on multiple pairs of facial images, tagged with corresponding source and target ages, showing the same synthetic and photorealistic person,” the researchers wrote in a related academic paper. By synthetically generating training data, they bypassed the “seemingly impossible task” of collecting images that “depict diverse identities, ages, and ethnicities from different perspectives.”

The result is what Disney calls a “production-ready” solution; this means it creates high quality output enough to be used in a real movie or TV show. Possibly the first AI solution of its kind that can dynamically change a player’s age in video despite varying expressions, lighting conditions and viewpoints. Researchers have also developed a user-friendly interface for FRAN that will allow artists to easily use the tool in a production environment.

Disney presented the research Wednesday in a paper titled “Production-Ready Face Re-Aging for Visual Effects,” and submitted it for inclusion at the 2022 ACM SIGGRAPH Asia conference in December. The paper’s authors include Gaspard Zoss, Prashanth Chandran, Eftychios Sifakis, Markus Gross, Paulo Gotardo and Derek Bradley, all affiliated with Disney Research Studios in Zurich, Switzerland.

Given Disney’s history of incorporating computer-generated actors into Star Wars movies and TV shows, including those who have been downsized using CGI, it wouldn’t surprise us to see technology similar to FRAN being widely used in future Disney productions. was announced.

Leave a Reply

Your email address will not be published. Required fields are marked *