low recognition quolity

  • Last Post 29 May 2014
Maksim Rodkin posted this 27 May 2014

Hi. I just setup the sample project and ran it. No one text was area recognised correctly, including printed text. Can i improve recognition accuracy? Please, see screenshot at URL link text

Order By: Standard | Newest | Votes
Maksim Rodkin posted this 28 May 2014

Hi all! i understood, the demo application has a bug. If you recognise entire document, you have possibility to check language, if you working with fields, you havn`t. This became a reason of my mistake in conclusion. Really, OCR is pretty fine.

Anastasia Galimova posted this 28 May 2014

Hi Maksim! It's not a bug, we just don't want to add all the parameters as a buttons. For field-level recognition language can easily be changed via the URL.

  • Liked by
  • Maksim Rodkin
Maksim Rodkin posted this 29 May 2014

Hi Anastasia, thanks a lot!