If something in the id or the image name could be matched up to to the chapter number, you could probably get a regexp to at least add that info to the title attribute. But even then, nothing's going to be able to extract the chapter "name" from the different images. As it is (with the image name starting at 9 and incrementing by twos), I don't think you're going to be able to do either... automatically. Your best bet is to add the ToC entry you prefer to the title attribute of each chapter-header manually.
|