Spectral color

This is what color night vision looks like now

Using deep learning AI, researchers at UC Irvine (go anteaters!) showed a proof of concept for color night vision.

Their AI was able to transform three images taken using only infrared light and colorize them pretty one close to the real colors of the scenes, Interesting engineering reports.

“You can imagine this having military applications, or just recreational applications to see at night,” said the study’s lead author, Andrew Browne, an engineer, surgeon and vision researcher at UCI. IEEE Spectrum.

“You can also imagine this use in chemical processing, where people have to work in places that are sensitive to visible light, or in medical or research situations where you have to work with light-sensitive retinal tissue.”

Using deep learning AI, researchers created a proof of concept for color night vision.

Unpublished websites: So you obviously can’t see in the dark, can you?

But it’s not still due to a total lack of light. The human eye can only see a narrow band of the light spectrum, from red to violet; wavelengths outside this range are invisible to us.

A room sealed with a lamp emitting infrared light would always appear black to the naked eye.

This is where infrared night vision comes in.

While our eyes cannot see outside of our visible light spectrum, technology can. Some night vision systems see in infrared light and output it as visible light, usually a single color.

This color often ends up as a sickly action movie green, because that’s the wavelength our eyes are most sensitive to, the Daily Beast explains. If you want night vision, it’s do not green, there are ultra-sensitive cameras that can glean and amplify visual light rather than infrared, but the use of visible light in science experiments can damage vulnerable tissue and specimens.

Browne Laboratory, UC Irvine Department of Ophthalmology

Towards fcolor night vision: “It started when I was sitting and thinking, ‘Is there a way to do everything in the dark? ‘” Browne told Spectrum.

The Irvine researchers turned to training a deep learning algorithm.

For their proof-of-concept study, published in the journal PLOS Onethe team started by taking a bunch of photos of color palettes and faces using monochromatic light sources, both with various visible and infrared lights, interesting engineering reports.

The AI ​​looked at all of these tagged images and learned to predict what an infrared image from a color night vision camera would look like.

To test it, the AI ​​showed 20 faces lit only with near-visible infrared light, Spectrum reported, which it successfully rendered in true visible light colors.

Work in progress : Even in this very narrow case, the results weren’t perfect – and crucially, the AI’s predictions are limited to what it was taught.

“There is some variability because you can put them side by side and see differences here and there,” Browne told The Daily Beast. “[But they’re] basically indistinguishable as if you didn’t even know you were looking at a predicted image.

Although far from ready for the battlefield, the technology could eventually have applications beyond the obvious military applications.

Although far from ready for the battlefield, the technology will eventually have applications beyond the obvious military applications, reports Spectrum.

“You can also imagine this use in chemical processing, where people have to work in places that are sensitive to visible light, or in medical or research situations where you have to work with light-sensitive retinal tissue.”

We would love to hear from you! If you have a comment about this article or have a tip for a future Freethink story, please email us at [email protected]