Hi,
I know that a lot of people who are interested in converting Arabic Subtitles (idx/sub) to text using OCR in SE through Tesseract, are diffidently familiar with the accuracy issues in Tesseract V. 3.x
Recently, I've encountered a person in Subscene who is claiming to successfully incorporate Tesseract V. 4.x within SE.
Tesseract V. 4.x has great improvement in regard of Arabic OCR accuracy.
I've made several attempts myself to incorporate Tesseract V. 4.x within SE, but in vain.
The issue is that the guy who did it is requesting money to teach the way of doing it and also to provide training files for each font.
You can see the videos here:
https://www.youtube.com/watch?v=NRh5njbC_V4
https://www.youtube.com/watch?v=wz8akEYraZw
https://www.youtube.com/watch?v=QEpbaKNEp0o
https://www.youtube.com/watch?v=mIffR8NOpxU
I would appreciate if someone know how incorporate Tesseract V. 4.x within SE and willing to share his way with us.
Thanks & Regards,
Mouna.