Startup Idea: OCR and Auto-Translation from Screenshot Tool

Startup idea to develop a software tool that allows independent scholars to automatically perform Optical Character Recognition (OCR) and translation on a selected screenshot from a text. This solves the problem of having to manually translate or run the entire document through an OCR tool when all that is needed is to understand a few lines.

I am an independent scholar who does historical research, and often I have to read texts in languages that I do not know so well. Often these texts are digitized, but they have not been subjected to OCR. (A lot of these were scanned from libraries by Google, I think, and put on the web.) The problem is that if there is no OCR, then you cannot do automatic translation. Of course there are products that can do OCR on a whole text, but this is a pain, since often I want a translation of a few lines. What would be good would be a way to take a screenshot of a rectangular region, and then automatically run it through OCR, and then automatically translate. I don't thinks this would be technically difficult, since basically it is just a matter of sticking together a bunch of technologies that already exist. Maybe I should do this myself. This is a pretty good idea. I either:

1. Put off reading the text. OR
2. Run the whole document through some online OCR.