Supasorn Suwajanakorn, YouTube screenshot
We call some people talking heads, but now they really are. Faces belonging to politicians and actors have been eerily reproduced in 3-D video, and these digital puppets can be made to say anything.
Want to see George W. Bush become a convincing Democrat? Or Ian McKellen adopt Daniel Craig’s persona in an interview? Expression transfers like this are a reality now.
The digital modeling was created by computer scientists and engineers at the University of Washington. It all started with the frequently photographed actor Tom Hanks. He’s played all kinds of roles over time, but he’s still easily recognizeable. Why?
Professor Ira Kemelmacher-Shlizerman, grad student Supasorn Suwajanakorn, and their team at the university wanted to capture Tom Hanks’ persona. Rather than taking turns reading lines from “Forest Gump,” they developed machine learning algorithms.
Their video (below) shows real Tom Hanks next to spot-on digital puppets of young Tom Hanks and current Tom Hanks.
“The technology relies on advances in 3-D face reconstruction, tracking, alignment, multi-texture modeling and puppeteering that have been developed over the last five years,” a university news article said.
Basically they pulled together more than 200 images of the actor’s face from Google Image searches, fed them to their algorithms, and out came a digital Tom Hanks model that completely captures his mannerisms and expressions — even if the words aren’t his, Mashable’s Lance Ulanoff explained.
After Hanks, the team threw other actors and politicians into the mix. Real video of George W. Bush starts out next to their digital model of him. Then Bush’s voice comes out of their Barack Obama digital model. And finally Bush gets parroted by a chorus of personae: Hillary Clinton, Arnold Schwarzenegger, Daniel Craig, and, yes, Tom Hanks.
The tech has even weirder potential. With enough photos, the technology could allow us to have realistic virtual reality interactions with people who aren’t even there. Historical figures, famous celebs, family members.
“You might one day be able to put on a pair of augmented reality glasses and there is a 3-D model of your mother on the couch,” Kemelmacher-Shlizerman told the university.