Be Unstoppable: Auto Image Captioning in Confluence

“A picture is worth a thousand words”. But does this old adage holds true for modern-day documentation and collaboration workspaces like confluence? We believe it is true. Here is a simple example, I want to tell people how important it is to take care of your car. On the left-hand side is a boring table containing the yearly values of the car, and on the right is a sleek graph. The conveys the same information as a graph on the right, but it is harder to read and somehow, even though it has more accurate data points, it conveys the message with less pizzaz. The graph on the right tells you clearly “if you don’t take care of your car, the car value value goes down drastically”.

Year Value of My car Value if i took care of my car 2015 45000 45000 2016 34200 44460 2017 29549 42237 2018 23639 38520 2019 19479 35496 2020 16050 34733 2021 11941 32996

But here is the accessibility conundrum. if the document only contains the graph, how does a screen reader read it? WCAG guidelines mention that images must contain alt-text, which can be used by assistive technology. However, enforcing a rule like that in a collaborative workspace like confluence is an administrative nightmare.

Unstoppable: Auto-Caption images

Addteq’s Unstoppable for Confluence plugin uses Azure Cognitive services to auto-generate captions for images. Automatic image captioning helps users access the important content in any image, be it a graph, chart, generic image or an image of text. The video below compares how images without a user-generated alt-text in confluence, would be treated by NVDA screen reader, with and without Unstoppable.