erdavids / Color-Cuber

Iteratively break down an image into sections of the average color underneath. Areas of greater detail are algorithmically prioritized and help to create an interesting effect.

Color Cuber

This is a simple Python program that utilizes the PILLOW library to iteratively break down an image into sections that slowly become more detailed. A section is colored as the average color from that p of the image. The program calculates a priority for the sections based on the area and error, making it so that more detailed portions of the original image will be more distinguished in the result.

This gif shows a summary of the process of the Color Cuber from 1 to 2048 iterations.

To use the program to create your own images, follow the process below (you will need to edit the file, indicated by #):

git clone https://github.com/erdavids/Color-Cuber
cd Color-Cuber

Place the image in the orig folder and edit the fields at the top for preferences

Python Color_Cuber.py

Examples:

I drew heavy inspiration from Quads by Michael Fogleman. It is an open source project that can be found . The only code that I used of his is the calculation of the average color and average error, although I edited the error function slightly. I borrowed the color_from_histogram function because I am unfamiliar with traversing histograms.