I’ve been using http://labs.ideeinc.com/multicolr/ to try and match images from the creative commons with segments of aaku. Sometimes with good results, other times not so great. Hopefully they’ll form a whole that provides a good visual summary of the chapters.
