I have seen some ebooks/papers that were apparently scanned from their paper versions but the text in the ebooks/papers can amazingly be copied out. I suppose the directly-scanned versions must have been processed by some Optical Character Recognition software.
So I would like to know what are the recommended Optical Character Recognition softwares? Especially those that are either for Ubuntu or free? If those for Windows are far more superior, please let me know as well.
I am particularly interested in those OCRs that can accept a scanned pdf file as input and still produce as output another pdf file that looks the same as the input one but with its text copyable.
Thanks and regards!
Please limit one software per answer