I have noticed that the "ara-Scheherazade" trained data was trained for the "Traditional Arabic" font. I have tried it, it performs well but with low accuracy, and has a problem when it comes to arabic numerals as the numbers are inverted. I want to fix the issue. I have tried to fine-tune it for it to better suit my data, but the fine-tuning is not working as it is also mentioned in the documentation that in order to fine-tune, I need to use the trained data found in the tess_data best repo. The main aim I'm trying to achieve is to manage to recognize both arabic letters and numbers. I know that there is a small issue with tesseract concerning both arabic letters and numbers, but the fact that the "ara-Scheherazade" font manages to recognize both but with a low accuracy means that it can be done, and I want to try and make it better. So does anyone know what can I do??
-- 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+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/4e9478bc-43a8-4b49-9891-c9e4b0ccccc8n%40googlegroups.com.
