OCR rotating scanned check copies and not recognizing text

Question

Hello,

I'm working with about 3000 check copies from bank records that were scanned in as a batch and want to make it a searchable document to isolate certain bank transactions. The scan quality is good and should be easy for OCR to grab. The images are of the front and back of the check so there is text that is running in different directions. An example image is below.

The checks are all printed in landscape, but were scanned in as portrait. I rotated them 90 degress in the PDF to be readable, but when I run the OCR it only recognizes about 50% of them and then it rotates the others back to portrait view. In some of the portrait view checks the OCR grabs the text that is perpendicular to what I am focusing on for this project.

Is there a way to force the OCR to focus on the text that is oriented the way I have rotated the pages?

Thank you for your help!

gary_sc · Answer

Hi, @Pat37368577ty5a, in a word, no.

One question is the front and back of the check one image or two? If it's one, you cannot get past this. Acrobat cannot deal successfully with two sets of text in two different angles. However, if you have text sideways when entered into Acrobat, it will see the direction of the text and correctly rotate it so it can be read. To keep the front and back together, do that by name. E.g., Check–1234a.jpg, and Check–1234b.jpg.

[Big hint: if you save the files as TIF files, when you open these in Acrobat, it will convert the files to PDF AND automatically do the OCR.]

Any written words can not be OCRed. Acrobat cannot understand written text.

I hope this helps

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