Each page would need to be scanned. You might just aswell keep the jpeg as an archive, what I would do is to simultaneously OCR it at the same time as scanning. Then compile a continuous text file. It wouldn't be so difficult, you would have a python script that would watch a folder and compile the text file. You can keep all the jpegs as a zip file, or a CBR file, which is a comic format, a zip archive of JPEGs are read in order. The other source file, will be a txt file, you can format however you want. word, OpenOffice equivalent, once you have worked on it, it would be easy to format as an ePub or pdf. Let me know if you approach any shop to do it, I think they would probably over charge - My missus could sit there and scan it or if you want to buy a cheap scanner, I can let you know how to proceed.