It was midday last Thursday when the founding team behind the new visualization app Pinscreen quietly launched what they hoped would be a welcome diversion from the tense last days of the election.
What Pinscreen says it has developed is a system for generating what can be photorealistic three-dimensional images from a single scan. While the technology’s current applications involved distorted faces for caricature, the same technology could be used for exceptionally realistic video.
Basically, it brings the technologies that were previously only available to folks at graphics studios like Industrial Light & Magic or Weta (the go-to studio for Peter Jackson’s Hobbit and Lord of the Rings series) to mainstream users.
“The idea is that you would normally need a highly-skilled digital artist to create this three-dimensional model. Now, you have a neural network that learns from the person and that does it,” founder Hao Li says.
The company had been in stealth mode for the past year, developing highly advanced artificial intelligence-powered visualization and mapping technologies to make easy tools for creating three-dimensional images, and they were finally ready to put the tech through its paces.
Their first targets? Hillary Clinton and Donald Trump.
Now as the polls close, users are getting their chances to use the Pinscreen technology and create some “comical” gifs to get their last licks at the two candidates — as the Los Angeles-based company gets ready for its next trick.
The company, launched by a core group of researchers from the University of Southern California, reveals why Los Angeles is becoming as much of a tech town as Tinseltown.
Founded by Li, the director of USC’s Vision and Graphics Lab at the Institute for Creative Technologies and a former lead researcher at Industrial Light & Magic, along with a team including Frances Chen and Shasta Ventures’ executive-in-residence Stanley Kim, Pinscreen is hoping to be on the cutting edge of three-dimensional imaging, with a technology that can create lifelike 3D images from nothing more than a snapshot.
A year ago, when Li was first putting the finishing touches on his research, a large social media company in the area came with an acqui-hire proposal to bring the Pinscreen technology in-house.
Li, seeing a sign in the interest from large corporations, decided to take the project to market independently.
So far, the company has raised $1.8 million in a seed round backed by Lux Capital and a few other undisclosed venture firms.
“I can take any video from the internet, any picture from the internet and mix and match things together,” to create a compelling three-dimensional image, says Li.
“In the end, we’re building a platform for people to create visual content,” says Li.
For now, that visual content is restricted to gifs making fun of the election and its candidates.
“We wanted to start this as an experiment,” Li said. “The hope is to boost the creation of gifs.”
Li’s technology currently focuses on facial scanning to create exaggerated images, but it can be more photorealistic.
“The most important thing is the face. It’s the most important biometric that you can identify,” Li says.
While faces come first, eventually the technology will get to an even more powerful point, he says. “In a year or two, we will be able to create things that are completely real,” according to Li.