A naive first attempt at conformaly mapping images. Given an image and an inverse map it will produce the image under the map. It works on animated gifs too.
Given an inverse map it scans over an empty image and for each pixel it finds the preimage poit and just copies over the value. This is why you need to supply an inverse map.
Just some examples of what it does to various images. (It likes tileable images a lot better).
It can map this 
To this 
It also does gifs! Mapping
To this
This program is based on python 3 and uses pillow to do all image manipulation.
Easiest way to set up would be to set up pipenv on your machine and simply run pipenv install in the directory with the pipfile as it will set up a virtual enviroment and install all needed packages for you. See here if you have no idea what I'm talking about.
Well theres a lot, kind of just threw this together but:
- It's slow. I'm doing the conformal map once per pixel, not sure if I missed some pillow primitive which might speed it up.
- It's inffecient with file size. The bilinear interpolation doesn't like palettes. if you give it a gif it will conver it to RGB. This takes up imense file size, pillow allows for some palette stuff, I just haven't done it yet.
- Branch cuts are sad. In general I think this isn't fixable, but in the case of the spiral function in conformal.py it definitely is. Just haven't sorted it out yet. Currently it makes spirals like this:
