pixelate turns your pictures into pixel art ! Well, sometimes.

It is a simple script based on PIL.

Algorithms will be added in the future. More precisely, I'd like to try to learn a mapping between input and pixelated space with a UNet-like encoder/decoder ConvNet.

It requires recent versions of both numpy and PIL.

pip install numpy

pip install Pillow

It was tested using Pillow 4.0.0 and numpy 1.12.1.

python3 main.py examples/jedsy_logo.png examples/test_1.png --num_color 10 --superpixel_size 10 --saturation 1.25 --contrast 1.2

num_color is the amount of colors wanted for the output. Small numbers typically give better results.

superpixel_size is the superpixel size. Rule of thumb : the larger the image, the larger the superpixels.

saturation is the saturation factor. Saturation helps create similar color zones.

contrast is the contrast factor. It is often useful to increase contrast to get better results.

new_size is the new size for the output image, three options available currently:

= 0. original = 1. superpixel size defines output image size = 2. user_defined image_size (x_dim, y_dim)

new_x_dim is the new pixel size in x dimension.

new_y_dim is the new pixel size in y dimension.

If the second argument refers to a folder, by default the name used for saving the processed file will be the same as the original file. An artifact is added if name refers to an existing file. ( B_edit: not tested above)

original work