A lot of our Confluence documentation contains images, often with formulas or graphs/charts/etc. Being able to extract information from images would be beneficial
