Ultra-Compact Metasurface Camera, the Size of a Grain of Salt, Takes High-Quality Full-Color Pics

Metasurface camera combines image capture with neural processing to blow past its competitors in image quality.

Researchers at Princeton University and the University of Washington have developed an ultra-compact camera sensor the size of a grain of salt — and say it can go blow-for-blow with sensors 500,000 bigger in volume for color image quality.

Easy to lose on the tip of a finger, the tiny camera sensor replaces traditional lenses with a metasurface made up of 1.6 million tiny cylindrical posts. Each post acts as an optical antenna, and with each post boasting a unique design work to shape the entire optical wavefront — passing the information to a signal processing system based on neural processing.

Combined, the two offer an impressive image quality: The team's testing showed that the sensor, working under natural light, not only beat previous laser-illuminated metasurface image sensors but could compete with a traditional sensor and lens system some 500,000 larger in volume - aside, the team admitted, from some blurring around the edges of the image.

"It's been a challenge to design and configure these little microstructures to do what you want," says study co-lead Ethan Tseng. "For this specific task of capturing large field of view RGB images, it’s challenging because there are millions of these little microstructures, and it’s not clear how to design them in an optimal way."

The solution: A simulator, which allowed for automated testing of different post structures, creating an approximation that provided the basis for fabrication of the silicon nitride metasurfaces — a process which, the team claims, could take place in a semiconductor fab and create sensors en masse at a lower cost than current methods.

The team's next trick: To boost the neural processing back-end, providing on-board computation for object detection and more — making the cameras ideal for use in everything from healthcare to robotics, the researchers say.

"We could turn individual surfaces into cameras that have ultra-high resolution," adds senior author Felix Heide of another possible area of interest, "so you wouldn’t need three cameras on the back of your phone any more, but the whole back of your phone would become one giant camera. We can think of completely different ways to build devices in the future."

The team's work has been published under open-access terms in the journal Nature Communications, while the source code is available on GitHub under the BSL-1 license.

Gareth Halfacree
Freelance journalist, technical author, hacker, tinkerer, erstwhile sysadmin. For hire: freelance@halfacree.co.uk.
Latest articles
Sponsored articles
Related articles
Latest articles
Read more
Related articles