Recommendation on how to best train Tesseract for new UTF-8 symbols

[Rafay Kalim]

Hey, so I am trying to train a new Tesseract model to only recognize certain UTF-8 symbols as I want an OCR that only recognizes these symbols and not other English letters etc. I realize there are two ways I can do this - one is to fine tune Tesseract over the normal English model and then blacklist the English text or train a completely new model that only recognizes this text. I was wondering if I could get some input into which of these - or another method, is better for ease, time and accuracy.

The context is I will have some various texts on a board and I want to recognize the locations of the symbols. However, I don't want to recognize any of the English or anything else as this may mess with my post processing. I have tried a few locations (like restricting where these symbols can be on the board and then only scanning the text in those strips) but I am not satisfied with the results. Additionally, I can also control the font and the size of the text on the board and everything else, except the actual codes. 

Thanks for the help!

[Lorenzo Bolzani]

Hi,

when you fine tune the model (maybe with ocrd-train) you can choose to restrict the model output to a smaller set of characters. No need to blacklist or anything else.

If you just want to locate the symbols something like opencv matchTemplate or training an opencv/dlib hog detector may be more appropriate. Using tesseract looks like a very convoluted way to do it.

If you have multiple symbols use multiple patterns/train multiple detectors.

Bye

Lorenzo

--
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 tesseract-oc...@googlegroups.com.
To post to this group, send email to tesser...@googlegroups.com.
Visit this group at https://groups.google.com/group/tesseract-ocr.
To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/3237ae86-db20-467c-bebc-6b45f854e799%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.