Simple way to get Recognition Server to ignore hand writing?

I would like to be able to setup Recognition Server to ignore handwriting and keep this as an image. This would be for documents that are forms and such with typed content and handwriting.

I have enabled the option under Output Format Settings > Format Settings "Replace uncertain words with images" but this does not seem to make any difference.

When scanning documents currently with handwriting the output will be that of character by character attempts at OCR leaving basically gibberish with some images of the handwriting intertwined - not a good look for a customer to present I can assure you.

Maybe there's something I need to set further somewhere or load a custom dictionary or something? I can't find what I need to be doing to achieve what should be a pretty common and basic requirement I would have thought.


Was this article helpful?

0 out of 0 found this helpful

Comments

3 comments

  • Avatar
    Sushi
    Wadeal,

    If you want to keep an area as image, you will need to setup an area template. Basically you can create one in the Verification Station. You set the area that are text, image, tables, etcs. Once you create those area, just save the area template. Then go back to the Workflow setting and add that area template into the Workflow setting. This all assumes that the documents you are working with are always going to be the same. Otherwise, you can setup a Verification Setup where an Operator can manually set the picture area as the document get processed.
    0
  • Avatar
    Vera Serova
    wadeal, there is no way to specifically exclude handwriting (other than using area templates as Sushi suggested, but if the handwriting appears in random places, it won't help much). What format are you converting the documents to, and what is the purpose of the conversion? If you are converting to searchable PDF, the garbage from the handwriting will be in the hidden text layer and won't be seen by the user. It also won't affect the search results because the meaningful information is still present in the file and will be found.
    0
  • Avatar
    killermax1
    thaaaaaaaaaaaaaaaaaaaaaaaaaanks
    0

Please sign in to leave a comment.