Hello.
Did Editor get a new OCR module in V7 as was announced earlier?
I did a quick check and saw the following results:
- OCR seems to use one CPU thread per page. So a single page only uses a single CPU thread, multiple pages use multiple threads.
- OCR still does not insert any "original white spaces" in between words of its own text layer.
- Automatic deskew can improve results in some parts of a part, but worsen results in others. I have one document where both Medium and deskewed Low/High turn "40qm" to 4q", only undeskewed Low/High results in the correct "40qm" being detected.
- OCR results differ between medium and high, but I don't know if high is supposed to always get better results now (in the past medium could beat high). Seeing how both Low and High could correctly detect the "40qm" example, I suspect that Medium still is "special"?!
OCR in V7
Moderators: PDF-XChange Support, Daniel - PDF-XChange, Chris - PDF-XChange, Sean - PDF-XChange, Paul - PDF-XChange, Vasyl - PDF-XChange, Ivan - Tracker Software, Stefan - PDF-XChange
-
Stefan - PDF-XChange
- Site Admin
- Posts: 19868
- Joined: Mon Jan 12, 2009 8:07 am
Re: OCR in V7
Hello Timur,
The new OCR module is still being worked on, but as soon as it is ready - we will make the switch - and simply installing the new build that includes the new OCR module will update it at your end as well. For the moment I do not have an exact date when this will happen - but we will announce it when we have such a date.
Regards,
Stefan
The new OCR module is still being worked on, but as soon as it is ready - we will make the switch - and simply installing the new build that includes the new OCR module will update it at your end as well. For the moment I do not have an exact date when this will happen - but we will announce it when we have such a date.
Regards,
Stefan