Israeli researchers unveiled a tiny component based on artificial intelligence that could enable individuals to conduct a variety of tests, with their devices — from checking blood oxygen levels to determining the amount of fat in a glass of milk and even lead to the develop of “neural cameras.”
Optical sensing devices have long been indispensable in medical and research settings, providing crucial information about the properties of various materials by analyzing the reflection or passage of light through them. However, the size and cost of traditional optical sensing devices have limited their accessibility to specialized laboratories.
But a team of Bar-Ilan University scientists led by Professor Doron Naveh have developed a component measuring just a few microns in size.
“Optical sensing devices provide information about the properties of a material through the reflection or passage of light through it and are used for medical and research purposes, but soon they could be part of our everyday life,” Naveh explained.
“Since the new component is only a few microns in size, it will be possible to integrate it into smart phones. We can examine the spectrum of anything we want, and share the information on social networks.”
This breakthrough, he said, could allow people to measure properties of food products, determine the exact color of objects, and assess their chemical composition and share the information — a concept he describes as “the spectrum of things.”
Said Naveh, “We can find out what’s in our glass, the percentage of fat in our milk, or whether the olive oil, honey, or juice we’re considering buying has been adulterated with scent extract.”
Although the development is currently in the laboratory testing phase, Neve and his partner, Prof. Panganian Xia from Yale University, plan to secure further investments to make their component a widely accessible product. The device was recently written up in the peer-reviewed Science journal.
Traditional optical sensing devices relied on the measurement of light characteristics, requiring larger equipment.
However, the article in Science explained that the Israeli device is based on “deep geometric sensing.” It replaces the optical instruments found in traditional sensors with an adaptive sensor, mathematical operations, and algorithms that can replicate the properties of light.
One key principle of “deep geometric sensing” is an “adaptive sensor” whose response to light can be altered through controlled variables such as voltage, current, magnetic fields, temperature, or mechanical pressure.
Data is obtained through training measurements, where the adaptive sensor is activated under various conditions, such as different spectrum ranges, temperatures, voltage levels, and pressures, resulting in a four-dimensional data space.
Decoding the information into spectral data is accomplished by an algorithm that mimics the functioning of a neural network.
Naveh suggested that this sensing technology could be extended to include not only the physical characteristics of light beams but also calculations within a camera equipped with hardware and algorithms resembling a network of neurons. This “neural camera” could analyze images, filling in missing details much like the human brain does.