Choosing background when generating output using PDF config.

[Jonas Winkler]

Hello.

I've got some input document input.pdf. This comes straight from a scanner and thus I do some preprocessing to improve accuracy (i.e., unpaper, black/white, increased contrast), which yields preprocessed.png.

When using the command

tesseract preprocessed.png output pdf

I receive a document, which has the ocr'ed text embedded. Great! However: Can I tell tesseract to use the original document input.pdf as the background (i.e., the one without preprocessing) of the generated PDF while still performing ocr on the preprocessed input?

Thanks,

Jonas

[Quan Nguyen]

I don't think Tesseract supports this. You may want to try to generate a text-only searchable PDF file and superimpose it on the original PDF file.