Description | Image |
---|---|
Target image | |
RGB MSE fitness | |
LAB MSE fitness | |
LAB Delta E CIE 2000 fitness | |
LAB Delta E CIE 2000 combined with SSIM fitness | |
SSIM fitness |
conda env create
Prepare a target image (store it in data/target_images/
) and a set of emojies/tiles (store them in data/emoji/
). Then run the following command:
python -m app.generator.generate
usage: generate.py [-h] [--starting-canvas STARTING_CANVAS] [--target TARGET]
[--fitness {RGBMSE,LABDeltaESSIM,LABMSE,LABDeltaE,SSIM}]
[-g NUM_GENERATIONS] [-p POPULATION_SIZE] [--width WIDTH]
[--height HEIGHT] [--emoji-size EMOJI_SIZE]
optional arguments:
-h, --help show this help message and exit
--starting-canvas STARTING_CANVAS
Path to an image to start with. If not specified, a
white image will be used.
--target TARGET Filename of target image. Should reside in
data/target_images/
--fitness {RGBMSE,LABDeltaESSIM,LABMSE,LABDeltaE,SSIM}
Choose fitness evaluator. See fitness.py for more
information.
-g NUM_GENERATIONS, --num-generations NUM_GENERATIONS
-p POPULATION_SIZE, --population-size POPULATION_SIZE
--width WIDTH If specified, resize the target image (and the
starting canvas) to this width. Otherwise, keep the
original width.
--height HEIGHT If specified, resize the target image (and the
starting canvas) to this height. Otherwise, keep the
original height.
--emoji-size EMOJI_SIZE