Quantum leap gives AI a sense of touch

Artificial intelligence can now “feel” and measure surfaces, thanks to quantum researchers at Stevens Institute of Technology.

Advances in computer vision and object recognition mean that AI systems can already “see,” in addition to their other human-like capabilities like holding a conversation. A sense of touch adds to the list.

A team led by Yong Meng Sua, a physics professor at Stevens, has developed a new method to give AI the ability to discern surface textures, such as distinguishing between a rough sheet of newspaper and a glossy magazine page.

The innovation, detailed in Applied Optics, combines quantum mechanics with AI. Sua and his colleagues, including CQSE Director Yuping Huang and doctoral researchers Daniel Tafone and Luke McEvoy, designed a system that uses a photon-firing scanning laser alongside advanced AI algorithms.

The system works by directing pulses of light at a surface. Reflected photons return carrying "speckle noise," typically considered a flaw in imaging. However, the team’s AI interprets these noise patterns to extract valuable surface texture data.

“This is a marriage of AI and quantum,” Tafone said.

In their experiments, researchers tested the method on industrial sandpapers with varying roughness levels, from 1 to 100 microns. Using mode-locked lasers, they achieved measurements with an initial error margin of 8 microns, which improved to within 4 microns – matching the precision of top industrial devices.

Get our latest stories today on Google News

Follow us

The approach was especially effective on fine-grained surfaces like diamond lapping film and aluminum oxide, Tafone noted.

Potential real-world applications include medicine, where AI could distinguish benign conditions from potentially life-threatening melanomas based on subtle differences in skin texture – an area where human examiners often make mistakes.

“Tiny differences in mole roughness, too small to see with the human eye but measurable with our system, could differentiate between these conditions,” explained Huang.

In manufacturing, the technology could improve quality control by identifying defects on a microscopic scale, ensuring components meet strict tolerances.

“Our method enhances existing technologies like LiDAR, already used in autonomous vehicles and smartphones, by adding the ability to measure surface properties at very small scales,” Huang said.