0

I am strugling with a problem of reading data from a table. If a number has got up to 2 decimal places everything looks OK, when number has got 3 decimal places I am not getting decimal point, ie. (instead of getting 5.575 I am getting 5575).

! enter image description here

I have also tried tesseract 3.02 from https://github.com/charlesw/tesseract but I found this less reliable in reading data from source I want to read, that is why I am using old tessnet2.

I am using following white char list: ocr.SetVariable("tessedit_char_whitelist", "0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ.,");

and language pack: english for version 2up.

Do you have any advices what can I modify in order to get more accurate values?

Thanks, Jack

Jack
  • 350
  • 2
  • 15

1 Answers1

0

Try bazaar or user-patterns in Tesseract 3.0x.

nguyenq
  • 8,212
  • 1
  • 16
  • 16
  • Thanks, bazaar is not a good option for me. As a solution to my problem I've found work around (deviation of the numbers from a reliable source canonot be greater that 0.3% so I can put some logic around numbers from tesseract). thanks for suggestion. – Jack Aug 05 '14 at 09:05