Hello everyone. I am trying to use Tesseract-OCR (v3.02) to recognise numbers (typically 5-7 digits).
The font used is *probably "*Futura Bold", but styled in a particular manner (see example). Using the *digits* parameter, Tesseract-OCR either gets it completely right, or fails to identify anything at all (return blank) depending on the crop (see example). Here are two examples where Tesseract-OCR fail: <https://lh3.googleusercontent.com/-5gzScT4G6Z8/VenBV8ABOLI/AAAAAAAAABQ/QnH_shWV7mI/s1600/atk.bmp> <https://lh3.googleusercontent.com/-V8lDEA4p3FA/VenBa9BxB8I/AAAAAAAAABY/CrRmh98bU9k/s1600/atk2.bmp> And is an example where it succeeds: <https://lh3.googleusercontent.com/-dTbycAo4o5Y/VenBicURHcI/AAAAAAAAABg/ZY4KU_77c2o/s1600/atk3.bmp> >From all my testing, as long as I do not forget the "digits" parameter, it will get it right if it gets it at all (without that parameter, it occasionally confuses 1 with l or I). And the tighter I make the box, the more likely it will succeed. The issue is that I am trying to automate the process as much as I can including the image capture step. Since the number can be anything between 5-7 digits I need to a bit of leeway (i.e. capture a slightly larger area). Perhaps with a bit more trial and error I might find a way to select a constant crop dimension that will work for 5-7 digits (most likely range I need), it would be preferable if I could get Tesseract-OCR to work with the second example. Is there anything I might be able to try on Tesseract-OCR side of things? I have only just started using this program. Thanks. -- 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/b0aaff28-4a88-4e34-8803-12d4fe6d426f%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
