You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@tika.apache.org by "Filippis, Ioannis" <i....@imperial.ac.uk> on 2013/08/14 13:59:54 UTC

pdf from scanner and header/footer/2-column cases

Hi all,

I wonder whether it is feasible to recognise whether the input pdf is a proper pdf (containing text) or a pdf from a scanner software (containing only images). For example in the second case I could use some ocr software like tesseract/tess4j. Looking at the Metadata.CONTENT_TYPE doesn't help, as it always equals to "application/pdf".

Moreover, I wonder how I could remove header/footer or even table text that are incorporated into output string and understand whether it is a two or one column pdf (so that I decide whether to sort by position or not).

Many thanks for your help.

Kind regards,
Ioannis