Being able to automatically describe the content of an
image using properly formed English sentences is a very
challenging task, but it could have great impact, for instance
by helping visually impaired people better understand the
content of images on the web. This task is significantly
harder, for