So I'm wondering what are the relevant publications one has to read in order to understand how some CAPTCHAs are hacked and how they can made more difficult to recognize by a machine. Of course there is basic image processing literature, but besides that? I'm also only interested in recent publication.
For example the first step that one has to do in order to recognize the text in an image is to segment it and find the location of the letters. So far I haven't seen captchas where the letters are moved so that they are connected to each other, which would make segmentation more difficult. Why is this?