2

I have seen some pdf files after OCR have very accurate text (both in terms of recognition rate, and in terms of alignment of text to the image).

What Unix/Linux tools help me tell, if the hidden text layer in a pdf file is from OCR or not?

closed as off-topic by Patrick, jasonwryan, jordanm, slm Dec 8 '14 at 13:49

  • This question does not appear to be about Unix or Linux within the scope defined in the help center.
If this question can be reworded to fit the rules in the help center, please edit the question.

  • 3
    This question appears to be off-topic because it is not about Unix or Linux. – Patrick Dec 7 '14 at 22:13
  • 3
    It is on topic, because it is about how to do that using applications running under Unix or Linux, not on other OSes. – Tim Dec 7 '14 at 22:18
  • @tim - I've removed that link since that's a copyrighted book and should not be shared here. Find another example. In the mean time I'm closing this as off-topic. This seems to be a better question for the tex site, potentially. – slm Dec 8 '14 at 13:48
4

The only images in that file are the front and back covers, the rest is normal text, not hidden behind a scanned image.

Another indication that this is typeset text and not OCR-ed is that special font characters are correct selectable despite their font and them being typeset as e.g. superset and also text in images where OCR normally gets confused.

Additionally pdfinfo indicates that the file is Created by DVIPSONE, i.e. that it is a (La)TeX file -> DVI -> PS and then converted by Distiller 4.05 for Windows to the PDF file. It is unlikely that a PDF file constructed from scans would have DVIPSONE as a creator

  • 1
    Another indication is that at 4.5 MB and 704 pages, the file is far too small in size to be an image-based PDF. Even highly compressed 1-bit images at reasonable resolution (200-300 dpi) are often 100 KB or more per page which would put us in the neighborhood of a 50-100 MB PDF. – a CVn Dec 8 '14 at 8:17
  • 1
    @MichaelKjörling That's right. You would more expect that from a DjVu file of a scan. – Anthon Dec 8 '14 at 8:19

Not the answer you're looking for? Browse other questions tagged or ask your own question.