Hi, I've followed the instructions on https://code.google.com/p/tesseract-ocr/wiki/TrainingTesseract3, and have successfully trained tesseract on a new language (ess). The results are OK, but I'd like to improve them.
My background is statistical machine translation, and from that context I would expect to be able to improve OCR quality through the use of character-level or word-level n-grams. Is there a mechanism whereby I can plug in an n-gram language model for my new language so that tesseract will use it? I've seen some references to bigram-dawg, but I haven't had any luck finding instructions for that feature, or even a good description of what it is. I may also look into dictionaries, but that will likely be problematic for me, because the language I'm working with is morphologically rich and highly agglutinative. Is there a mechanism for providing a morpheme dictionary? Thanks, Lane -- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/tesseract-ocr. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/b82468f3-b8f8-433d-aabd-78043d12919c%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
