Despite all the technology at hand, computer-generated imagery of humans can still look a little ropey, with teeth perhaps proving the greatest challenge when it comes to accuracy.
If it’s a digital double you’re after, the process usually requires diving into the subject’s mouth with a scanner, or time-consuming tooth-by-tooth modeling work.
But now a team of boffins at Disney Research appears to have come up with an impressive solution that’s notable for its speed, simplicity, and most importantly, accuracy.
Its remarkable technique uses just a few snaps of a person’s pearly whites, or a simple video taken on nothing more than a smartphone. Even better, the subject need do nothing more than say “cheese.”
Working with ETH Zurich and the Max Planck Institute for Informatics, the Disney team developed its technology by first creating a model of a “regular” set of teeth by combining data from high-resolution 3D scans of 86 sets of teeth.
This gave the software a deep understanding of the shape, color, and spacing of human teeth, and helped researchers to design an algorithm that could accurately create a copy of a person’s entire set of teeth simply by analysing the real thing in several photos or a video.
Incredibly, the algorithm can also build an accurate reproduction even if parts of the teeth are hidden from view in the visual material it’s working off.
Imaging teeth is ordinarily a tricky task, but as the video above shows, the researchers appear to have hit upon a method that produces remarkably accurate results.
“Our algorithm only requires minimal user interaction and can operate on a set of individual, uncalibrated images, making teeth capture as easy and convenient as taking a few pictures or even a short video clip using a standard mobile phone,” the team said in its report on the technology.
“Image-based reconstructions of the human face have grown increasingly sophisticated and digital humans have become ubiquitous in everyday life,” Disney Research VP Markus Gross told EurekaAlert.
Gross added that by combining creativity and innovation, the research “continues Disney’s rich legacy of leveraging technology to enhance the tools and systems used to create more realistic and believable digital actors for films or video games,” with the medical industry another sector likely to find the technology useful.
- A.I. perfectly predicted last year’s Super Bowl score. What happens to betting?
- This A.I. literally reads your mind to re-create images of the faces you see
- Social (Net)Work: What can A.I. catch — and where does it fail miserably?
- The best speakers
- Deep learning vs. machine learning: what's the difference between the two?