Facebook has updated the Automatic Alt Text (AAT) system that generates automated photo descriptions for visually impaired people. According to Facebook, AAT can now identify a photo 10 times more than before and identify more detail in a photo. Taking to their blog, Facebook wrote that users who are blind or visually impaired (BVI) could experience Facebook imagery, provided it is tagged properly, with the help of alternative text.
A screen reader can describe the images’ contents using a synthetic voice, enabling them to understand images on their Facebook feed.
Automatic alternative text (AAT) was introduced by Facebook in 2016 and was recognized in 2018 with the Helen Keller Achievement Award from the American Foundation for the Blind.
As per Facebook, AAT “utilizes object recognition to generate descriptions of photos on demand so that blind or visually impaired individuals can more fully enjoy their News Feed.”
Now, AAT has been upgraded to represent multiple technological advances that improve our users’ photo experience.
Facebook has expanded the number of concepts that AAT can reliably detect and identify in a photo by more than 10x. Descriptions are also more detailed, with the ability to identify activities, landmarks, types of animals, etc.
Facebook also made it possible to include information about the positional location and relative size of elements in a photo. Thus Facebook is now able to specify the exact number of figures in an image.
Facebook has explained that AAT uses simple phrasing for its default description, rather than long and flowy sentences. It is highly functional, and users can read and understand the descriptions quickly.