Facebook creates software for visually impaired … – Dynamo

as the Internet is full of images, Facebook will launch a system that you can “read” the images and thus communicate its content to those with a visual impairment.

Launched Tuesday, Facebook is about to change the way you access your images. Through Artificial Intelligence servers the most popular social network can decode and describe the images uploaded to this site, so that they can be read.

>

the person behind this innovation is Matt King, an engineer working on Facebook and who lost his sight as a result of retinitis pigmentosa, a condition that destroys the light-sensitive cells located in the retina.

“most things that happen on Facebook are extremely visual,” King says. “As a visually impaired person, you feel very out of topics of conversation.”

the technology that King and his team developed software uses Facebook to recognize objects, in order to decipher what it contains an image. so far, the program can recognize things like food and vehicles.

“Our artificial intelligence has advanced to the point where we find it convenient to get computers to describe images in a meaningful way,” added the engineer.

> Although in its infancy, the program “helps us move in the right direction regarding the inclusion of everyone who wants to participate in the discussion and feel part of the team,” adds King.

currently, the software is able to describe basic things like: “There are two people in this picture and smile.” Anyway, from the social network ensure that the program is trained to recognize about 80 common objects, from trains to cars and meals, through mountainous landscapes and beaches, and sports such as tennis, swimming and golf. All this information appears in the “alternate text” for each photo. The greater the number of analyzed images, the greater the degree of sophistication in the software.