problem with font sizes

hi all.. i've come across this issue and cant really figure out a way to resolve it. i'm trying to write a small program that will process multiple images through tesseract and look for a particular sequence of characters. the images will probably come from all kinds of sources - scans, fotos, et

not words

hi all... is there a way to instruct tesseract to ignore anything that is not trained to read. like the lines around the date and time in this image: http://quereven.com/images/moo_time.jpg the simple reason is that with those lines the output is mostly garbage. if i just "white" out the lines t

Re: not words

that was awesome. thanks... On Apr 15, 5:10 pm, namenick wrote: > hi all... > > is there a way to instruct tesseract to ignore anything that is not > trained to read. like the lines around the date and time in this > image:http://quereven.com/images/moo_time.jpg > > the