This project's goal was to make an image compressor using the "kMeans" algorithm, fully in haskell using clustering. This was made on Fedora linux using Haskell and stack to build.
Each colors of the image are grouped into "Clusters", then, based on the number of colors you want, it chooses the most prevalent colors to keep, making a compressed image.
How to compile:
make
You can test it doing this:
./ic_tools_fedora36/convertImg your_image > converted_img.txt
./imageCompressor -n NbColors -l convergenceNb (for example 0.08) -f converted_img.txt > compressed_img.txt
./ic_tools_fedora36/xpmImg compressed_img.txt > compressed.xpm
And Voilà ! You have a compressed image :D
Example with 4 colors:
It was a fun project to do, a bit hard because doing this type of algorithm in Haskell is hard, but I like the result !