IIRC, "Omnipage" will do OCR from a scanned pdf page; just be prepared to do some proofreading; the OCR can be a little iffy, depending on the qaulity of the original scan. Also, if the book has been published, perhaps there is a better quality scan somewhere on the 'net. I once found a public domain book on the Project Guttenberg site (
www.guttenberg.org) that was scanned into a formatted Word doc...