Microsoft Edge now provides users with automatically generated image descriptions.
In a blog post last week Microsoft’s Travis Leithead announced a new feature in Edge (Via The Verge). The feature provides its users with automatically generated image descriptions when a website does not provide them. Edge is the browser that replaced Internet Explorer on Windows and is available for most devices. Currently, it’s available on Windows, macOS, and Linux, but not on iOS or Android.
An image description (or alt text) is a short description of an image for assistive technology such as screen readers. They ensure the content or meaning of these images is accessible to those who can not see them.
Image descriptions are too often forgotten on websites. This feature aims to fill those gaps and provide a better browsing experience. We’ve previously seen similar technology on Facebook where they call it automatic alt text. As on Facebook, Edge uses AI to understand and translate the contents of the image to a description. Currently, it supports most file formats and 5 languages for its descriptions.
Images are processed in the cloud by the Vision API, so users need to opt-in first. You can do this by enabling a setting labeled “Get image descriptions from Microsoft for screen readers” in Edge. It skips very small or large images, images that are purely decorative, or images containing mature content.
Don’t forget your descriptions!
For now, this feature will not yet improve your experience on Twitter, because Twitter substitutes empty descriptions with “Image”. It is one area where this feature can improve in the future.
The blog post and the video rightfully remark that while this feature is useful it can’t replace well-crafted image descriptions. To tell automatically generated descriptions apart, the Vision API starts them with “Appears to be” or “Appears to say”. It goes on to provide some resources to help make better descriptions.