Hi, This thread is pretty old, but trying my luck here. Were you able to successfully read the MICR data from bank checks? I need to extract account and routing number from images of checks and was looking for a way to do that. There were a couple of commercial libraries which I tested and they work good. But if we can get an open-source one, that'll be awesome. I tried using tesseract, but it is not able to do accurate OCR on MICR data. Were you able to train it to read the numbers correctly?
~ Anurag On Friday, March 28, 2008 4:44:59 AM UTC-4, ttutuncu wrote: > > Hi, > > I am working on a project in our company. My aim is to recognize MICR > 13B-type fonts which are generaly used for number printing on bank > cheques. > > I have managed to successfuly read the bank cheque number. The success > rate is around %95 if I only cut the region of the number from the > scanned image. > > The problem is when I process the whole image (the whole bank cheque) > tesseract tries to recognize all the regions and and does not want to > ignore characters that do not occur in the allowed unicharset. > > My unicharset is: A,B,C,D,1,2,3,4,5,6,7,8,9,0 > > What I want tesseract to do is to ignore every other character that is > not in the unicharset. > > When tesseract tries to assimilate every character in the image, the > success rate of recognizing the cheque number decreases to around %70- > %80. > > Is there a way to make tesseract ignore the other characters that are > not in the unicharset. > > The other problem is that in the output I sometimes get a "o" > character instead of a "0" (zero) character even though it is not in > my unicharset. > > Does the DangAmbigs file really work? Because I think it is not > working for me. Is there a configuration for this? > > When I finish my project I will release it to this group. > > Thank you for your help... > -- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/tesseract-ocr. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/12c879df-5d1f-4dd4-ac30-fbb8aa322739%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
