I saw it in action at hackMIT and it seemed to work really well.
You can use something like Onenote to save the image, which will OCR the image and make it searchable and also gives an option to Copy text on right click.
Disclaimer: Microsoftie
I'm working on an OCR project at the moment; I'd love to see which library they used.
He used Tesseract for the OCR, I don't recall what he used to make it highlightable. It was quite impressive when he demoed it on a XKCD comic.