This archived news story is available only for your personal, non-commercial use. Information in the story may be outdated or superseded by additional information. Reading or replaying the story in its archived form does not constitute a republication of the story.
SALT LAKE CITY — Cameras are designed for humans, but one University of Utah professor believes it might be time to start designing them more for machines.
Electrical and computer engineering associate professor Rajesh Menon and his team of researchers engineered a camera that can take pictures through a regular pane of glass or window, rather than a curved lens.
If a normal digital camera sensor is pointed at something without a lens, the resulting picture just looks like a pixelated blob. That’s not much use for the average Joe; but in that mess is enough information for a computer program to properly detect and identify an object. It just needs an algorithm to decode it.
So Menon and his team of researchers set up an inexpensive digital camera sensor and connected it to the side of a plexiglass window with reflective tape wrapped along the edges. They then pointed it toward an LED light board positioned at a 90-degree angle in front of the sensor.
With the help of a computer running the algorithm, the camera was able to capture recognizable images and videos from the light board. Thanks to the reflective tape, the light from the LED board passed through the glass and about 1 percent of it scattered through the window and into the camera sensor so the computer could decode the image.
“We have always thought of cameras as human-centric … (but) in the future, and even now, most images or video may not be seen by humans; they may just be seen by machines,” Menon said.
Homeowners with security cameras, for example, won’t see every frame of footage, but a machine can detect certain objects and alert the homeowner.
Menon’s design could make it possible to construct a home with security cameras built into the windows or a car with a dash cam ingrained in each pane of glass. The technology could even be used to help autonomous cars sense more objects surrounding the car.
“It’s not science fiction, but it would require some engineering work,” Menon said.
More immediate applications for the design are most likely in the world of augmented reality, he added. Current AR glasses have cameras that are pointed at the wearer’s eyes to track their positions. But with Menon’s technology, those cameras could be positioned on the sides of the lens and could drastically reduce the size of the glasses.
“It’s not a one-size-fits-all solution, but it opens up an interesting way to think about imaging systems,” Menon said. “Why don’t we think from the ground up to design cameras that are optimized for machines and not humans? That’s my philosophical point.”
Menon and his team will continue developing the technology to include 3-D images, higher color resolution and photographing objects in regular household light.
A paper detailing his full experiment was published in "Optics Express."