Hi to All! I try to train tesseract to "computer-like" and "digital-like" fonts (rugsnatcher-demo and levi-adobe-dia in particular). I kept the training recommendations for image preprocessing, changing box file using qt-box-editor, spaces between characters. But despite these I got fails when I put the trainign command (e.g. tesseract eng.rugsnatcher_demo.exp0.tif eng.rugsnatcher_demo.exp0 box.train). In some cases I got through but finally at combine_tessdata procedure (Error: traineddata file must contain at least (a unicharset fileand inttemp) OR an lstm file.).
Editing box files seems totally useless to me since changing 1 "failing" character (that is correct btw) makes even more fails during training. I have already tried another font sf_digital_readouts. Considering this, training was successful only for few characters, but when I put there paragraph of text I got some Fails as well. I am stuck at this few weeks but I am not able to make good results for this. I am curious what am I missing? Is it possible to teach tesseract these fonts or should I take another approach? Thank you for any response! -- 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 https://groups.google.com/group/tesseract-ocr. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/de9d3d76-115b-42f8-a80a-806588d85059%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
