Skip to main content
aric43249394
Participant
July 19, 2016
Question

Accurately OCR a scanned form to get input data

  • July 19, 2016
  • 2 replies
  • 715 views

Hi, so I am trying to accurately OCR a populated form in order to get the input data (ideally to do this for multiple forms). When I simply OCR the form Adobe does not recognize accurately where the tables and inputs are. I have a soft copy of the form that I'd like Adobe to somehow use when OCR'ing the populated scanned version of the form.

Please help!

This topic has been closed for replies.

2 replies

Lovekesh Garg
Adobe Employee
Adobe Employee
July 20, 2016

OCR can recognize only text of table not the tabular view.

If you are facing any issue in text accuracy, please share a sample document. We are trying our best to make the text recognition system better every time. You can use https://cloud.acrobat.com/send  for sharing the file.

Thanks.

CtDave
Participating Frequently
July 19, 2016

OCR evaluates the picture (scanner output image) for possible images of characters. If such are recognized the software makes a 'best-guess' as to what the character might be. While OCR can be rather accurate it is not 100%. Many variables can impact recognition accuracy.  As well, for "tables", "columns" & such -- these are an arbitrary human interpretive artifact. An image/picture has none of these and OCR only recognizes potential characters of text.