Now with OpenCV Backend!!

I re-wrote the whole thing in C++ and statically linked it to OpenCV (originally) in order to use just a few upscaling/preprocessing things, but then that got me thinking:

This would probably be doable™ entirely in OpenCV, and that'd give me a lot more control over the backend

So uh, I did that.

A lot of the same frontend bugs still exist, but it's now easier to install, since the builds produce statically linked binaries.

v0.1.0 - It... functions... Ship it?

It's 2AM, and I've been pecking at getting the latest version of Autotrace to work with this plugin for weeeeeeeeks. I just merged it, and honestly, I feel like it's enough to cut a v0.1 release. I've demoed it, and I've been working really hard.

So I'm doing this before I change my mind.

🚨 THIS IS A PRE-RELEASE! THINGS ARE BROKEN, THIS IS NOT EASY FOR GRANDMA TO USE YET. But it's a cool proof-of-concept. One that is very, VERY unpolished. This isn't the code that I took to Makerfaire (though in hindsight I probably should have cut a release of some kind for that as well), but it's like... the first step...?

Anyway, don't hurt yourself. Please open an issue if you find something interesting.