Jump to content
  • 0

(Archived) OCR questions

Tyler-new to evernote


How does one determine if their PDF has been OCR'd by Evernote? An Evernote employee stated:

"You can retrieve the text recognition results from the resource.recognition field. If the field is unset, then recognition hasn't been performed yet or won't be performed (e.g. if the resource isn't an image)."

....but I have no idea where the "resource.recognition" field is. A simple icon demonstrating that 1) that OCR has been performed on the document, 2) that OCR cannot be performed on the document, or 3) the document is in queue for OCR seems like a no brainer to me for Evernote developers!

Also, another larger issue regarding OCR:

If a PDF document is originally scanned sideways (unreadable) and I rotate it so that I can read it (in say, Apple's Preview application), once OCR is performed by Evernote, is it performed on the ORIGINAL sideways document or the altered right side up one? I ask because I signed up as a premium user so that I could benefit from Evernote's OCR capabilities, but when I search for a word that I know to be within a PDF that was rotated, my search results identify that the word is in the document as I would expect it to, but the highlights are COMPLETELY off making the OCR function basically useless.

Thanks in advance for comments/answers! :)

Link to comment

1 reply to this idea

Recommended Posts

  • Level 5*

If you check to see whether there's a little magnifying-glass-on-a-book icon in the top right margin of your note, the colour should tell you whether a note has been OCR'd. Since it's a small icon and I have trouble seeing any colour in it, I tend to right-click a PDF file in the note to see whether there's a "save searchable PDF" option. If there is, the file has been OCR'd.

An easier way to confirm OCR is to do it yourself - which would answer the second query too. We've only just got that nice highlighted search terms feature, and while I don't know a great deal about the process, I do know that an OCR'd file has a "hidden" text-only version parked behind the picture you see of the page. The text version will likely remain unchanged if you mess with the orientation of the pages - hence you get a mismatch if you change and save the file without repeating the OCR.

In any event I don't see how you can say the OCR is useless - it finds you the document and even the right page!

Link to comment


This topic is now archived and is closed to further replies.

  • Create New...