If not, you will have to get rid of all the unnecessary line breaks with Tesseract, but the actual character recognition is better than with Ocropus.
In my tests tesseract was almost 100% accurate, except it missed a few spaces, and a few symmetrical apostrophes ' were turned into left-hand side single quotes ‘
This is much better than the OCR engine included in Microsoft Office, and MUCH better than gocr
1. Install from here (512 kb)
2. Copy everything from /local to /usr/local, then you can delete /local (I made a mistake packaging it... I'll package version 3 sometime and get it right).
2. Download your language file (English, French, German, Dutch, Spanish, Italian, Portuguese, Vietnamese or Old German) from http://code.google.com/p/tesseract-ocr/downloads/list (900 to 1400 kb)
YOU WANT THE "LANGUAGE DATA" file, not the "SOURCE TRAINING DATA" file. If you have a different language, you'd better make them
3. Extract the language file into /usr/local/share (so the stuff should be ending up in /usr/local/share/tessdata).
4. Run with e.g.
Code: Select all
tesseract /path/scan.tif /path/output_file
It will automatically append a .txt extension to the output.
It ONLY works with uncompressed and G3 compressed tiffs because I disabled libtiff support because of a bug that they tell me will be fixed in the next version. Xnview, nconvert, Imagemagick convert, and probably other things can make these. I'm guessing Xsane does too. The Gimp can't (or at least couldn't )
FYI I had compile problems with 2.03, so we're waiting for 2.04