Detection of lightly rotated documents

[cherry]

Hi Olivier,

Thanks for your sharing your discovery and solution with us.

[OlivierM]

Hello,

I tried to find a page that i could send you, but they are all confidential.

But good news, I found the problem:
It must be some security on the pdf!

When i print to pdf one of my pdf, I can OCR with foxit and it works great on all documents.
So thanks for the help, and we will certainly buy a few licences for our group.

Olivier

[cherry]

Hi OlivierM,

Is it possible to export the page that you thought been skipped during the OCR process? Without the file, it may not be easy to further investigate the issue. Thank you.

[OlivierM]

I just thought:
If the pdf was created with adobe acrobat, is it possible that with some security features, Foxit cannot OCR the pdf file?

[OlivierM]

Hello,

No unforntunately, I cannot provide the pdf which poses a problem to me since it's confidential information.
But I artificially made another one (using word), which i scanned badly on purpose more rotated than my confidential document.
For some reason, the OCR performed very well and detected all words on all pages.

Since I cannot provide my original confidential pdf, could you help me and evaluate what could make the OCR to not detect the words?
Apart from the rotation, all documents I have to scan were originally made on Microsoft Word and scan quality is descent (words are very slightly blur, but still easily readable).
At the end of the process I still get the message "they are page(s) which don't have editable text, they are scaneed or image-based"

When the OCR starts, he takes a few seconds for the first page then continues to the next pages but the process is much faster for all pages after the first one. So he really seems to skip those pages.
I tried on a few others pdf obtained the same way (archived scanned documents to pdf). On some non rotated file, he skips also the pages. Sometimes, he doesn't even find words on the first page, skipping all pages

Thanks

Olivier

[cherry]

Hi OlivierM,

Is it possible to provide the mentioned lightly rotated scanned PDF document for testing? You may upload it here or email it to [email protected].(Attn: Cherry) Thank you.

[OlivierM]

Hello,

I'm looking for a solution to OCR already scanned documents
Foxit Phantom standard 6.2 seemed a nice program and it seems to be efficient.

Except:
When I OCR a document that was slightly rotated during scanning process, the OCR does not OCR the page at all, skipping it.
Is there a feature to auto detect slight rotation in documents? Or is there any version (business?) that does the job better?

Thanks
Olivier