The social network launches on Tuesday a system that describes the content of the photographs uploaded by users. . A commitment to accessibility, which ensures that Facebook will continue
“This image may contain: pizza, food.” The description may seem limited, but for those who can not see is an important step.
Facebook decided to launch a new service that allows blind people to know what they contain images that are placed on the social network by users.
currently, there are software for the blind ( screen readers ) that make computers accessible through the reading on the screen. But these systems can only read text. At a time when two billion images are uploaded every day on social networks, Facebook has found a failure.
Through Artificial Intelligence, the company decided to develop the “Automatic Alternative Text”, a system decoding and describes the images and allows systems reader screen read
in the video you can see how the system works:.
the system it is available from Tuesday to Facebook users that use screen readers on iOS set to the English system, but will be extended soon to other languages and platforms.
Matt King, engineer of this social network is one of those responsible for this new tool. “On Facebook, a lot of what happens is extremely visual,” King said to the BBC , “a blind man can feel that is put aside in conversation.”
The engineer knows well than speech. He lost his sight due to retinitis pigmentosa, a rare degenerative disease. For King, this new system is an achievement, a path to equal rights.
“I feel I have the right to this information. I’m asking something that is already available to others also reach me. so I think it is a matter of justice. “
for now, the system is still very basic. The software can recognize 80 familiar images, such as cars, trains, food, mountains, water, beach or some sports. But the more photographs enter the system and are interpreted, more sophisticated will be.
The goal of Facebook is evolving in the area of accessibility. . In the case of images, for example, use visual recognition to identify people in a photo and say the name
The team provides:. This is only the beginning
No comments:
Post a Comment