just too difficult for me !

92 views

Skip to first unread message

Méric JJ

unread,

Feb 4, 2024, 6:30:54 AM2/4/24

to tesseract-ocr

HI all,

A few years ago I have managed to train Tesseract for Bamanan (african language also called bambara). Then suddenly, versions changed, neural networks were introduced, and all my work became unusable.

I tried several time to read the documentation with the intention of starting all over again, but it's just too complex and involved (for me)...

Could someone help me?

Bamanan is a latin language, it just adds 4 new letters to the alphabet used in this message (and their capital variants): ɛ ɔ ɲ ŋ .

I can provide text for training based from real world text.

I can provide lists of frequents words and dictionaries (I am the author of spell-checker for bamanan).