Shreeshrii / tessdata_arabic

Finetuned traineddata files for Arabic
29 stars 7 forks source link

tessdata_arabic

Finetuned traineddata files adding support for numerals and punctuation in Arabic script

PlusMinus Finetune Training was done based on tessdata_best/script/Arabic.traineddata by tesstrain.sh using fonts and training text for approximately 4000 iterations.

ara-Amiri.traineddata Info

combine_tessdata -d ara-Amiri.traineddata

ara-Scheherazade.traineddata Info

combine_tessdata -d ara-Scheherazade.traineddata

ara-Scheherazade.traineddata Info

Fonts used for plus-minus training

'Amiri' \
'Sakkal Majalla' \
'Scheherazade' \
'Traditional Arabic' \