large char set language training

[Richard Foo]

Dear all,

I am new to tesseract. When I train a large char set language like Chinese, I have no idea which step I should use the char set(over 7000 char) I prepared. Currently, I consider it as a training set by converting all_char.txt to tiff files. Therefore, I have a image training data of a single font which can be used for making box files. 

p.s: is there any methods(softwares) for rendering synthetic training data from text except scanning or printing? 

thanks,

Richard

[ShreeDevi Kumar]

Yes, there is a method for rendering synthetic training data from training_text and fonts via text2image program and tesstrain.sh script.

https://github.com/tesseract-ocr/tesseract/wiki/TrainingTesseract

https://github.com/tesseract-ocr/tesseract/wiki/Training-Tesseract-%E2%80%93-tesstrain.sh

Which version of tesseract are you using?

I would suggest that you try the latest version built from github with the Chinese traineddata and then do the training.

ShreeDevi
____________________________________________________________
भजन - कीर्तन - आरती @ http://bhajans.ramparivar.com

--
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-ocr+unsubscribe@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/5aee4e28-4bb4-460a-8d27-c9ff3a8a3bd0%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.