I have found a wonderful utility that converts pdf over to txt / html and allows you to edit sections of it before the pdf is converted over.
It is called PDFMasher,
http://www.hardcoded.net/pdfmasher/
You load in a pdf, and it displays each line of text, you can sort by columns, which is handy to group things like page numbers from a scanned book, and mark them to be ignored when you generate the output file.
I have no association with the creator / website etc for pdfsmasher, just a very happy user.
I would process your pdfs first using this program, then load them into word / bookcreator to do final editing / clean up etc.
It works wonders for me.