We have seen a shift of online materials from written words to an increasingly picture, gif and media dominated space and while this is a good thing for aspiring photographers and users who find it easier to be engaged via this means, it leaves out visually impaired users for the most part and therefore presents an uneven playing field.
Despite the availability of varying assistive technologies which visually impaired Facebook users normally employ to use the social network, they are unable to connect and interact; as is the norm this days, with visual imagery.
Facebook is however set to remedy this to an extent with a new service which it is set to launch on Tuesday. The sophisticated navigation technology which blind people use to make a computer usable is called a screen reader and they basically turn the contents of the screen into speech output or braille but they are unable to read images.
Good News For The Visually Impaired Facebook Users
The Facebook servers will now however engage an AI which can decode and describe images uploaded to the site, providing them in a format which can be read out by the screenreader to the visually impaired Facebook users.
The product was developed by Matt King a Facebook engineer who lost his sight as a result of retinitis pigmentosa which is a condition that destroys light sensitive cells in the retina. Working alongside his team, the new tool was developed using Facebook’s in-house object-recognition software to decipher what an image contains and has already been trained to recognize items like food and vehicles.
Currently, the system describes the images in a fairly basic way, like stating the amount of people contained in a picture and what they are doing but Facebook says it has now trained its software to recognise about 80 familiar objects, from cars and trains, to food and settings such as mountain, water and beach, and sports such as tennis, swimming and golf. It adds the descriptions as alternative text, or alt text, on each photo. The more images it scans, the more sophisticated the software will become.
A similar function was added by Twitter last month, the difference being that Twitter enables users to manually add their descriptions of the images; which may make the descriptions better, but the user has to actively choose to do it.