OCR scanned document

Hi there,

I am testing scanning documents using the exactscan plugin and when the document hits DEVONthink, pages that are oriented in landscape are rotated and cut off during the OCR process. If I just scan the document, it doesn’t rotate the pages automatically, it seems to happen during the OCR phase. Is this expected behaviour?


Looks like the simple answer is to make sure the PDF is oriented correctly BEFORE OCR. Then everything works as expected. This document was a mix of portrait and landscape, so I guess that just doesn’t work correctly.

1 Like

Yeah, in my experience you want to have any other manipulation done before OCR. I’ve had issues scanning books and realized I just needed to split the pages so they’re one book page per pdf page instead of facing pages, prior to running the OCR, as doing it afterwards gave some strange results.

1 Like

That’s correct. We are working with the OCR developers to see if we can get this worked out properly. I obviosuly can’t give any timeframes or promises, but it is under investigation.