Science

AI with vision superior to the human eye

You will also be interested


[EN VIDÉO] Playlist: all about transhumanism, robotics and artificial intelligence
Jean-Claude Heudin, director of the IIM (Institute for the Internet and Multimedia), tells us about his vision for transhumanism, the future of artificial intelligence and robots in this series of interviews, compiled as a playlist.

Researchers at the University of Central Florida have unveiled a new artificial intelligence capable of seeing, recognizing shapes and identifying objects. This technology can be used in robotics or to improve autonomous vehicle systems. Will vision become a new sense that will soon be open to artificial intelligence?

In any case, this is the project of his researchers at the University of Central Florida. With a device capable of playing retina one eye human, their creation can lead to the creation of a new, more efficient AI with new capabilities. In particular, technology may allow this artificial intelligence instantly understand what she actually has ahead of her. The most obvious use case seems to be autonomous car or robotics.

Another advantage is the technology described in a recent study published in review of ACS Nanoalso works better than the eye regarding range wavelengthsthat is, maybe understand as well asultraviolet what light ordinary. For unmanned vehicles, the versatility of the device should enable safer driving in a variety of conditions, explained Molla Manjurul Islam, lead author of the study.

Universal Device

“If you are in your self-driving car night and the car’s imaging system only works at a certain wavelength, say the visible wavelength, it won’t see what’s in front of it, Molla Islam describes in the study. But in our case, with our device, he can really see the whole thing.” According to the researchers, there is currently no such device that can work equally well in the ultraviolet range, the visible wavelength range, and even infrared. These skills give this device a unique character.

Currently existing intelligent visualization technologies work in several stages, from information discovery to its processing, including storage. “We had devices that behaved like synapses from brain person, but we didn’t provide them with the image directly,” says Tanya Roy, a professor at the University of Central Florida. Now, adding to them the ability to perceive images, we have synapse-like devices that act as “smart pixels“in-camera by simultaneously detecting, processing and recognizing images.”

In addition, the device does not need to be multiplied, and the chip is very compact, about the size of a thumb. At the moment, the accuracy level is about 70-80% for each wavelength, and it must continue to improve before it can be used. The researchers estimate that the technology could be ready within the next five years.

Interested in what you just read?

.

Back to top button

Adblock Detected

Please consider supporting us by disabling your ad blocker.