Hello Everyone,
I am doing Read PDF with OCR activity as my pdf is in Image so I want to convert it but not getting a proper output :(. If anyone please help me out !
Regards,
Hemal
Hi
OCR is never reliable and therefore we can’t expect 100% correct response. Licensed OCR engines which primarily run on Cloud like Google Cloud OCR or Abby Cloud are better in terms of efficiency.
However do note that they are paid.
2 Likes
yes, true. So what do we do? What if there is handwritten text?
Hello Hemal,
From my perspective OCR is just a last resort option when in comes to automation. In my own tests only 3% of the PDF where scanned successfully. If you have more than one PDF do your math (haha). Therefor I would suggest (depending on how much PDF’s) to copy paste the content into a word document and read it out of that one.
Have a great day!
Handwritten text cannot be processed by OCR as they work well only for electronic texts. You can try looking for any suitable ICR engine (Intelligent Character Recognition).
However I haven’t worked on it yet.
For ICR try parascript. I have not integrated with RPA before but was successful in BPM applications.
It works best when there is something like a table on confined set of parameters to check address, such as Address Database, list of possible values. Names , email etc are very difficult as there are infinity possibilities.