Add character to existing font in Tesseract training

Question

Is there a way to add a character to an existing font that was trained for Tesseract OCR?
I just trained a new font, but it is misreading colons (':') as '2.'
So, I would like to just train it to recognize the colon, without having to retrain everything I just trained. Is that possible?

Answer 1

Incremental training in tesseract is not possible (adding some letters to existing data).