Tesseract OCR finds too few boxes / ignores small characters -
i have problem training/text recognition process tesseract. here trainingdata: http://s11.postimg.org/867aq10ur/dot_dotmatrixfont_exp0.png while training tesseract ignores dashes (i've marked them red boxes, make clear ones mean) , if i'm using trained data text recognition ignores them. today i've played around tesseract parameters (setvariable(name, value)) unfortunately had no success. can teach tesseract dashes? thank in advance!
tesserect training pretty tricky.
your best chance might handle dashes single char. if box editor or whatever tools using not see dashes all, try running image processing first, threshold or invert. try taking @ opencv. have excellent tool kind of image processing.
Comments
Post a Comment