Artificial intelligence with a vision superior to the human eye

Researchers at the University of Central Florida have unveiled a new artificial intelligence capable of seeing and recognizing shapes and identifying objects. This technology can be used in or improved robotics Autonomous Vehicle Systems. So will sight be a new sense that will soon be unlocked for artificial intelligence?

It is anyway a project of her researchers from the University of Central Florida. Thanks to a device capable of reproduction Retina from one eye Human, its creation could lead to a new and more efficient artificial intelligence with new capabilities. Concretely, technology can allow this artificial intelligence To immediately understand what she really has in front of her. The most obvious use case seems to be a file independent car or robots.

Another benefit is the technology, described in a recent study published in reconsidering ACS nanoalso results in better eye performance with respect to range wavelengthsthis can imagine feeling as wellultraviolet that light Normal. For self-driving vehicles, the versatility of the device should provide safer driving in a variety of conditions, explained Mullah Munjer Islam, lead author of the study.

Versatile device

” If you are in your own car the night And the car imaging system only works on a certain wavelength, to transmit the visible wavelength, it will not see what is in front of itMullah Islam describes in the study. But in our case, with our devices, he can actually see whole According to the researchers, there is currently no device of this type that is able to work well in the ultraviolet, that is, the visible wave band and even Infrared. These skills give this device a unique character.

Existing intelligent imaging technologies operate on several stages, from information discovery to information processing, including storage. ” We had devices that acted like clamps The brain Humans, but we did not provide them with the image directlyUniversity of Central Florida Professor Tanya Roy explains. Now, by adding image sensing capability to them, we have clip-like devices that work like “smart pixelsIn the camera by detecting, processing and recognizing images at the same time “.

In addition, the device does not need to be swiped, and the slide is very small, about the size of a thumb. Currently, the accuracy is around 70-80% per wavelength and must continue to improve before it can actually be used. Researchers estimate that the technology could be ready within the next five years.