numerical anal.

...because if I called it numerical analysis, no one would click ;)

Anyway, my thesis, like, works 'n shit. Like, I did a bunch of calculations by hand (errrrg) and ran the same input through my program and it spat out the right answer.

As many of you know, I'm a big proponent of open source and I release all of my code. However, I'd kinda prefer not to release this as a standalone. It's definitely not sufficient for facial recognition on its own, but it's very valuable to any program that uses a correlation template. IOW, oftentimes a facial recognition program will compute a difference measurement between the input image and candidate images in a database. This sounds simple but the notion of "difference" is actually quite unclear. My program provides one such definition. While it's not really robust enough on its own, it can be used in say, the training of a neural net. That's why I don't want to just code a GUI and release it as a facial reognition program; I'd rather put it out into the wild and see what people do with it.

So my question is, does anyone here know of an open source library for either numerical analysis or image processing that might want this? GSL already rejected it because it's too specific.

BTW, for those who don't know, my program computes affine-invariant Fourier descriptors for the polygons resulting from a segmented image. I also have a vector class and a segmentation program (I wrote both), and I use GSL for complex numbers.