People rely on pictures and graphics to share information and experiences, but not everyone can see these images. For the blind and visually impaired, this can be very challenging if image descriptions aren’t provided for their screen readers and braille displays to take advantage of.
Google is addressing this problem. Today, it is announcing a new desktop Chrome visual accessibility feature that will help people more easily navigate the web with a little help from machine learning.
There are millions of unlabeled images across the internet. When you come across one of these images, screen readers and braille displays will describe them as “image,” “unlabeled graphic,” or even the original file name, which is often an unhelpful string of numbers.
Google is using the same technology in Google Lens and Google Photos to offer more accurate image context for those who need it. The machine learning technology will try to recognize the unlabeled image and provide a better description.
For example, if a user comes across an unlabeled picture of a cat, the tool might say “appears to be a cat lying on a couch.” The “appears to be” tells the user that this new Chrome accessibility feature is doing its best to provide visual context.
This feature can also read the text in images. After processing something like a receipt or social graphic, the tool might say “appears to say” and then go on to read the content. These qualifiers help alleviate confusion users might experience when coming across these computer-generated descriptions.
The feature is new, and translations aren’t perfect, but Google is working diligently to enhance the visual accessibility features on Chrome. The tool has labeled more than 10 million images so far, and it is continually improving.
If the tool isn’t confident about what the image content is, it won’t provide a description. Whether the images are labeled or not, the content won’t be shared with web admins or developers, even though humans would be able to generate better descriptions.
To take advantage of the new Chrome visual accessibility tool, go to Settings, then Advanced, and in the”Accessibility” section, users can enable “Get image descriptions from Google.” This can also be enabled for each web page by right-clicking and selecting “Get Image Descriptions from Google” in the context menu.