color-harmony.py

A python script that extract main color groups from an image using k-means clustering and returns the color harmony based on the RGB wheel

Dependencies

• Python 3.7+
• sklearn
• Pillow
• matplotlib

Installation

Pip

The stable releases of gallery-dl are distributed on PyPI and can be easily installed or upgraded using pip:

$ python -m pip install color-harmony

From Source

Clone the repository or download the folder from github, then navigate to the respective directory and run:

python setup.py install

Usage

To use color-harmony, simply include the path to your image file:

$ color-harmony [OPTION]... [FilePATH]

See also python color-harmony --help.

Examples

Default run on an image file will return both RGB colors and create a color palette image with the same name but with "_colors.png":

$ color-harmony sample.png

You can also set the number of clusters, tolerance, and output configurations:

$ color-harmony -k 5 -t 1.5 -o text icon.png

Color Scheme Analysis of Popular Illustrations (Python Notebook on Github)

You will need to download the dataset from: https://www.kaggle.com/profnote/pixiv-popular-illustrations or use your own images.

ColorSchemeAnalysis.ipynb

Main jupyter notebook for the analysis, can use the data files below:

harmonies.csv

File containing the harmonies of each illustration in the dataset.
Column names: "Monochromatic", "Complementary", "Split Complementary", "Triad", "Square", "Rectangular", "Analogous", "Other"

wheelColors_arr.csv

Wheel color representations of each illustration in the dataset, starting from Red -> Green -> Blue -> Red.