Layout of text does not match graphic

[guebert]

Hallo,

first of all many thanks for this great feature! The OCR function is the only reason for me to have Acrobat 5 still in use.

The result of the new OCR function seems to be "wrong sized" , because the text does not match the graphic. Please have a look at the files attached.

(Same problem is reported here: https://forum.pdf-xchange.com/ ... 35&t=12327)

Many thanks,

Michael

[Paul - Tracker Supp]

Hi guebert,

see my answer on that other post you referred to. Try rasterizing the entire page during the OCR process using "Convert Page Content to Image only - Add Text as layer". Doing this on your page resulted in correct placement of the OCR test layer. See attachment.

hth

[guebert]

Unfortunately this option produced a much bigger file (2,5MB) than the PDF out of my scanner (350 kB), so it's not a good solution.

Please post, when the layout problem was really fixed.

Many thanks,

Michael

[Jamie - Tracker Supp]

Hi guebert,

We will have our OCR developer take a look when he returns from vacation next week.

Regards,
Jamie.

[mhill]

Ditto. At first, the OCR does not appear to have any function at all--the first thing I tried was to search for some clearly readable text. When I found another forum post that notes that the text is available for copying and pasting, I discovered that text was indeed there in the bottom left corner of my page. OK, that's useful for indexing pdfs and locating the page where text appears. Then I found your suggestion about converting to an image. That indeed works to align the text to the image, but the image size an order of magnitude larger than the fax image files I'm testing with, even at 150dpi.

I bet that your developers could find a scale factor to fix and get it to work with the Preserve Original Content option.

[Paul - Tracker Supp]

Hi mhill and guebert,

your points are taken, and we will be looking into this.

Do remember that this is a free offering - not a paid for product and it is really only intended as a way to make documents searchable and/or copy paste into other documents.

Regardless, we do appreciate the feedback and will endeavour to make this better in the coming weeks and months.

hth

[Walter-Tracker Supp]

This is a bug. I will work with the relevant Viewer OCR developer team members to ensure that this is corrected. Thank you for bringing it to our attention.

-Walter

[Walter-Tracker Supp]

Just an update: we have fixed this issue in our development version and so you can expect it repaired with the next release!

Thanks again for bringing it to our attention.

-Walter

[guebert]

Just an update: we have fixed this issue in our development version and so you can expect it repaired with the next release!

Many thanks! I know, "it's ready, when it's ready", but maybe you can tell me, if it will be this month or ...

Michael

[Tracker Supp-Stefan]

Hi Michael,

I don't have an exact date for the next build, but we do usually release them about 6-8 weeks apart now with most of the team focused on ver3.

Best,
Stefan

[enaef]

Hi

First - A big thank also from me I am very happy about this feature . I work with the pro version and I certainly would be willing to pay for an advanced OCR functionality!
One feature I wish is the ability to OCR documents which have mixed horizontal and vertical orientation.

I had the same problem (text noch matching graphic) this morning and found with this thread that I'm not the only one. Thanks for fixing it in the next release. (Besides inexing and being able to copy past I also am happy to being able to use the text tools (such as highlighting) when OCR works correct).

Ernst

[Tracker Supp-Stefan]

Thank you for the kind words Ernst!

You can be sure we are working on improving our OCR engine as much as possible!

Cheers,
Stefan

[enaef]

Hi
The problem regarding the layout text not matching the graphic seems really to be fixed with the new .201 release.

Thanks a lot

Ernst

[Tracker Supp-Stefan]

Hi Ernst,

And thanks for the update - glad to hear that!

Cheers,
Stefan