Thank you for all your replies.
There are some 'guidelines' or some kind of 'best practice' to determine hidden text/element in a pdf file?
I mean, if Acrobat is able to determine 'hidden' text using heuristics, can I try to reproduce a something similar using SDK?