Disney has a high-quality AI tool that can automatically make actors look younger or older

The Disney team has created a cool new AI tool that can automatically make actors look younger or older in real time, in high quality. I know we’ve seen filters like this on social media apps, but this next-level technology is ready for production. I have to say I’m impressed and I’m curious to see filmmakers start using it for real.

I love learning about new film technologies, and here are all the details that have been shared with a video that you can watch below:

Photorealistic digital reaging of faces in video is becoming increasingly common in entertainment and advertising. But the predominant 2D painting workflow often requires frame-by-frame handwork that can take days to accomplish, even by experienced artists. While facial image reaging research has attempted to automate and address this issue, current techniques are of little practical use as they typically suffer from loss of facial identity, poor resolution, and choppy results in subsequent video frames.

In this paper, we present the first practical, fully automatic, production-ready method for reaging faces in video images. Our first key insight is in tackling the problem of collecting longitudinal training data to learn how to re-age faces over long periods of time, a task that is nearly impossible for large numbers of real people to accomplish. We show how such a longitudinal dataset can be constructed by exploiting the current state-of-the-art in facial reaging which, although it fails on real images, provides photorealistic reaging results on synthetic faces.

Our second key insight is therefore to leverage such synthetic data and formulate facial re-aging as a practical image-to-image translation task that can be performed by training a well-understood U-Net architecture, without the need for network designs more complex. We demonstrate how the simple U-Net, surprisingly, allows us to advance the state of the art for re-aging of real faces on video, with unprecedented temporal stability and preservation of facial identity across expressions, points of view and varying lighting conditions.

Finally, our new FRAN (Face Re-Aging Network) incorporates simple and intuitive mechanisms that provide artists with localized control and creative freedom to direct and refine the re-aging effect, a feature that is very important in actual production pipelines and often overlooked in related research work.

Watch the video and let us know what you think!