How to train tesseract to recognize more complex characters?
Guys, can anyone faced with the task of training the tesseract? I would very much like to add such functionality to your program, but can not understand. As I imagine it: in tesseract enters the picture the words on the entrance, he recognizes, I correct if necessary, and thus it updates its dictionaries tessdata in the future is already correctly recognize this word! At the moment, the very recognition of it is poor...Slipped binarybonsai the text and the paragraphs and lines and words and even characters, the way the characters turned out to be the most uzhastnoe recognition, for some reason... And here is my last hope for learning
I will be glad to any Council!