Smart glass concept recognizes images on its own with analog computing
Researchers at the University of Wisconsin-Madison have developed a “smart glass” concept that can recognize images without a power source, sensor, or processor, and could potentially be used for facial biometrics in the future, the University News outlet reports.
The team, which also included a researcher from the Massachusetts Institute of Technology, published their proof-of-concept research in a paper on “Nanophotonic media for artificial neural inference” in the Photonics Research journal.
“We’re using optics to condense the normal setup of cameras, sensors and deep neural networks into a single piece of thin glass,” explains UW-Madison Electrical and Computer Engineering Professor Zongfu Yu.
Yu says the wave dynamics of light propagation enable the new method of analog artificial neural computing.
The concept involves segmented glass with embedded bubbles and impurities strategically placed to bend light in particular ways to differentiate images. For the proof of concept, the researchers made glass pieces that could recognize written numbers well enough to detect when a ‘3’ was changed to an ‘8’ in real-time, according to the report. The concept is similar to machine learning, but with training on an analog material, rather than digital code.
“We could potentially use the glass as a biometric lock, tuned to recognize only one person’s face” says Yu. “Once built, it would last forever without needing power or internet, meaning it could keep something safe for you even after thousands of years.”
The researcher’s work is supported by a grant from the Defense Advanced Research Projects Agency (DARPA).