This repository is a collection of image-processing techniques mostly developed in order to learn about methods and/or test things. As such, most of the files here do not have a functioning CLI.
There are currently two implementations of wavefunction collapse. One is image-based tile-by-tile in prelim.py. The other numpy-based and is able to collapse things in many dimensions and not just images. This has both the pixel-by-pixel and tile-by-tile methods, and is in wavecollapse.py. These both have the same basic CLI:
python3 prelim.py --orig <path to input image> --out <path to output file> --w <width of output> --h <height of output> --tw <tile width> --th <tile height>
The wavecollapse version requires an additional flag --alg <pixelwise|tilewise> which is whether to do a pixelwise or a tilewise collapse. The input should be a .png image, although probably other formats will work. The output will be saved as the image format you specify. Check PIL docs for the list of accepted formats. In the case of prelim, contradiction pixels will appear as black. For wavecollapse, you can specify the contradiction pixels when using as a library, but from the command line these pixels will also be black. The pixelwise implementation is still somewhat buggy.
The Wave Collapse.ipynb ipython notebook shows how to use the implementation in wavecollapse as a library.
Edge detection and various image processing techniques that rely on edge detection. This program does not have a CLI and contains several methods that go directly from input image to output image.
Dithering using Bayer matrix. Converts the image to black and white.
Uses k-means clustering to reduce the total number of colors in the image.
Uses a median cut to reduce the total number of colors in the image
Various methods for interacting with pixel color and images, including BFS and pixel distance metrics.