OCR and exporting to Word problem

[XChangeGirl]

I have a pdf that consists images of plain text. The origins of the images was simple screen shots of formatted text, as you would see in a typical world document. Nothing special, just paragraphs and occasional subheads.

When I OCR these in pages in pdfxce, everything works as expected. All the text is recognized and is searchable.

But when I export it to MS Word, the exported docx document is very weird. Each page has what appears to be a transparent image (or other object) overlay that prevents text from being selected, and when that image or object is selected and deleted, all the text goes with it. The document is only about 10,000 words and would normally be under 1Mb in size but the exported Word doc is almost 8Mb with all the extraneous stuff in it.

I can copy the text straight from pdfxce as either rich text or my pasting it as plain text, but the results are unsatisfactory for various reasons.

I've tried different settings in the export to Word dialog, but none of them are making a useful difference.

Any ideas what's going on, and why what seems like a simple OCR procedure is producing these results?

Thank you!

[Stefan - PDF-XChange]

Hello XChangeGirl,

Are you using the standard or Enhanced OCR engine?
If you are using the standard engine - the text that you can search inside the PDF file is invisible and on top of the original image.
If you use the EOCR - it should be able to replace the letters from the image with actual visible letters inside the PDF file and the conversion to Word should be much better.

This KB article should help you get better results:
How do I edit document text after OCR has been performed?


Kind regards,
Stefan

[XChangeGirl]

Thank you. I had been using the standard OCR rather than the enhanced scan. That seemed to make a difference.

[Paul - PDF-XChange]