Millions of colors and shades can now be distinguished by machine vision

(ORDO NEWS) — Researchers at Northeastern University have created a device that can recognize “millions of colors”.

Scientists have developed a machine vision system that captures the entire color spectrum, and a neural network that recognizes colors. The accuracy of color perception in the new system is higher than in humans.

A person distinguishes about a million colors. But this is its natural limit. He couldn’t see better. The same cannot be said about machine vision, which is constantly being improved.

Researchers at Northeastern University have created a device that can recognize “millions of colors”.

Scientists have developed a machine vision device they call “A-Eye”. It is able to analyze and process colors much more accurately than a person.

Co-author Swastik Kar explains why high color resolution is so important: “In the world of automation, shapes and colors are the most commonly used elements by which a machine can recognize objects.”

Millions of colors and shades can now be distinguished by machine vision 2
Schematic diagram of A-Eye operation

When you point a digital camera at a colored object and take a picture, the light from the object passes through a set of detectors with filters that differentiate it into primary colors – RGB (red, green, blue).

“You can think of these color filters as funnels that funnel visual information into individual blocks and assign a specific number to natural colors,” Kar says.

RGB filters impose a limitation on the perception of color, but a person does not see them: he has only three color detectors in his eyes. And what if there are not 3 such detectors, but, say, 33 or more?

Kar and his team used “transparent windows” made of a unique two-dimensional material that has a lot of such “funnel” in it.

“Our device recognizes colors differently,” Kar says. “Instead of breaking it down into its main components, we use all the spectral information.

We use some methods to change and encode the color. And as a result, we get a set of numbers that help us recognize the original color much more accurately than with the usual three-color coding.

The trained neural network is engaged in changing and coding colors.

The research has many industrial applications, including drones, crop sorting, remote satellite imagery, and even trash sorting, Kara says.

Online:

Contact us: [email protected]

Our Standards, Terms of Use: Standard Terms And Conditions.