Implemented K-Means algorithm from scratch in C++ for compressing a RGB image into another image having the K most dominant colours. Used opencv for Image reading and processing.
Make sure you have opencv C++ installed, then run,
g++ KMeansCompressv2.cc -o output `pkg-config --libs opencv`
./output rain_princess.png rain_princess_compressed.png 16
Command line arguments are <input_file_path> <output_file_path> <n color vectors> .
PNG format recommended, JPEG actually increases file size in some cases(which I think is due to the nature of the format and the fact that we are saving with 100% quality).
