Improvements to OCR

Question

I would like to know if Adobe has incorporated any new AI capabilities into any newer versions of Acrobat Pro to improve optical character recognition. My company currently uses Acrobat Pro 2017, and similar to many others we have frequently experienced issues with OCR being inaccurate unless the PDF files are pristine with perfectly legible text. Improvements to OCR (with capabilities more similar to ABBYY Finereader) would be a dramatic improvement for customers. My company frequently works with older (clear) hand written documents which I have successfully OCR'd with relatively minimal user input in ABBYY Finereader 15, as well as tables of data which would we would like to export to Excel for re-creation, but Acrobat Pro 2017 will very rarely work even on scanned printed text documents unless they are near perfect 600dpi scans with no skewing. Also, the lack of ability to change or correct the erroneous OCR results currently makes the feature far less useful than it can and should be.

Abambo · Answer

I'm using the Acrobat OCR on all of my scanned documents. It's not perfect, but it delivers a searchable file. However, it is not a full fledged OCR reader as a third party specialized product.

I do not remember the performance of the Acrobat Pro 2017 in detail, so I cannot tell you the difference of the OCR engine. I simply can't remember of major problems, when my scan has been done correctly. Scans from third parties very often have the issue, that they are not done with the OCR functionality in mind. And very often, I even have problems reading the text myself.

As said, it is not a full fledged OCR and correcting the results is not that easy. If you want to reconstruct many documents to be able to rework the content in Word, I would still go with an external OCR program.

Sign up

To post, reply, or follow discussions, please sign in with your Adobe ID.

Sign in to Adobe Community

To post, reply, or follow discussions, please sign in with your Adobe ID.

Scanning file for viruses.

This file cannot be downloaded