News 

The Stanford University engineers are developing a "multi-aperture image sensor" that not only captures an ordinary 2D image, but also creates a "depth map" of every object in the frame.

The sensor contains thousands of tiny lenses, each mounted on an array of 256 pixels. The pixels have been shrunk to just 0.7 microns each to allow them to cram so many lenses on to the one sensor.

Every item in the photo is captured by at least four of the chip's micro lenses, producing

ADVERTISEMENT

overlapping views from slightly different perspectives - similar to the way human eyes work.

"It's like having a lot of cameras on a single chip," claims Keith Fife, a graduate student working on the project. "In fact, if the prototype 3-megapixel chip had all its micro lenses in place, they would add up to 12,616 'cameras'."

The depth-sensing camera could potentially be used for facial recognition, with the lenses able to pinpoint the distance to the person's nose, ears and other facial features.

It could also find applications in 3D modelling, architecture or biological imaging. "You can choose to do things with that image that you weren't able to do with the regular 2D image," Fife says. "You can say, 'I want to see only the objects at this distance,' and suddenly they'll appear for you. And you can wipe away everything else."

But perhaps the most intriguing use is for improving the "eyesight" of robots, giving them better spatial vision than human beings. ""People are coming up with many things they might do with this," Fife adds.