Setup

Build the docker container, this also builds the code:

docker build -t dcq:latest .

Use the run_all.py script to batch process multiple images:

docker run -it dcq:latest
python3 run_all.py -i <input_path> -o <output_path> -p <palette_size> -m <dither_mode> 

There are more options available:

Options:
  -i, --input_dir PATH
  -o, --output_dir PATH
  -k, --kernel_size INTEGER       Specify the kernel size used by the DCQ
                                  approach.
  -p, --palette_number INTEGER    Specify the maximum number of colors used
                                  for quantization
  -d, --dcq_path PATH             Specify the path to the DCQ executable.
  -m, --mode [DCQ|FS|FS-ICM|FS-ICM-DCQ|FS-DCQ]
                                  Specify which dithering algorithm to use.
  -a, --alpha_mode [divide|quantize|none]
                                  Specify how the program should handle the
                                  alpha channel.
  -c, --cluster_mode [k_means|median_cuts]
                                  Specify the clustering algorithm to
                                  initialize the palette for Floyd Steinberg
                                  Dithering
  --help                          Show this message and exit.

Code Structure

algorithm.cpp

Contains the core functionality for the dithered color quantization approach.

alpha.cpp

Contains code to handle alpha channel pre- and post-processing.

floyd_steinberg.cpp

Contains the dithering code for the floyd steinberg algorith.

init.cpp

Contains codee for initializing all parameters for the DCQ approach.

PBar.cpp

Simple progress bar code to track time passed and current progress.

utils.cpp

Small utility functions used in multiple locations throughout the code.