Good thought, however touch-ups done in graphic programs are usually done to an existing element. In other words, if these were touch-ups, the text would most likely also exist in the main text layer. A close watching of the following video shows this is NOT the case:
http://www.youtube.com/watch?v=EgVIei87oFo
There’s another important issue here which has not been addressed. The text is curved at the appropriate place at left. This makes no sense whatsoever for a document “constructed” by somebody typing in text. Who would first type in text, put it in a book, photograph it, then mix that curved graphic image with other (typed in) letters? Ignoring the enormous needless effort you would spend doing things, you couldn’t possibly get a good result that way.
It just makes no sense. No, the right explanation is the simple one: the layers were created by a software program trying to make sense of, and do OCR on, a scanned image.
notice the lines not even a trace of evidence anyplace on the paper. If that were possible I would have to say that is the best program ever made.