I really like the idea. If your vision is to expand NI Vision which is already a pretty good base, I would be interested in the added functionalities. We have been doing Vision projects in the past and among other features, the lack of a better OCR has been a pain. I agree with ShaunR that supporting such a tool is the main challenge over time. If the project gets off the ground and we can share some of our functionalities, I'd be willing to do so.
Good luck!