CuneiForm Batch subprogram - available in the English-interface distribution http://www.cuneiform.ru/downloads/setup ... rm_eng.exe - produces optionally the so-called FED-files (that have the *.FED extension).
Screenshot:
Choose "NewShortcut5" - it's the link to the Batch subprogram in the CuneiForm package. (weird name??? some developer's mistake probably, in Russian version it's named "Batch OCR").
When you open the Batch subprogram, create a new batch:
Setting the new batch, choose "FED files" at some point:
Run your Batch OCR task. After the end you will see the FED-files in the output folder.
The FED format contains the raw OCR information.
I have translated the original FED format documentation ( http://www.cuneiform.ru/downloads/doc.zip , subfolder "ced") from Russian to English. It is available on my site here:
The description of the ED-format (written 2.11.1998)
http://www.djvu-soft.narod.ru/ed_discr_en.htm
The description of the ED-format ver.2000 (written 16.09.1999)
http://www.djvu-soft.narod.ru/new_ed_en.htm
The idea is to create a some kind FED-> djvused-txt converter.
See also the djvused doc: http://djvu.sourceforge.net/doc/man/djvused.html#lbAV (paragraph "Hidden text syntax")
Who could do that?