Facebook launches a new technology that improves the so-called 'AAT', the Automatic Alternative Text.
This helps blind and visually impaired people to understand the content of the photos posted on the social network.
First introduced in 2016, the AAT uses computer vision technology to automatically generate descriptions of an image.
When using screen readers, technology converts text and other elements into speech.
In the most recent update, Facebook allows AAT to recognize content over 10 times more specific in Instagram and Facebook posts than in the first development, five years ago.
As a result, more photos will get the text description and more spoken text detail for better understanding.
"We have expanded the categories of existing items and content and added new ones," explains Facebook.
The new technology has the ability to recognize activities, symbols and types of animals, to have more in-depth descriptions.
So we will not only have phrases like "Selfie outdoors" but "Selfie of two people outdoors, with the Leaning Tower of Pisa".
Another novelty is the ability to provide detailed information on a specific image, including the position of the objects in the photo (top / center / bottom or left / center / right) and their relative size (primary, secondary).
The descriptions are formulated in a simple way, which facilitates translation into 45 different languages.
The AAT is available on Facebook for photos featured in NewsFeed, profiles and groups.
Furthermore, it can also be used when a photo is opened in "detail" mode.
On Instagram, on the other hand, the technology works on photos in the Feed, Explore and profiles.
The detailed description of the image is accessible by pressing and holding on the photo for Android devices and via "custom action" on iOS.