Assessing Image Compliance with the 60-30-10 Colour Rule

Quantifying Colour Balance in Images Using Clustering Algorithms.

Last updated on October 03, 2026

The 60-30-10 rule is a design principle used to create balanced, cohesive colour schemes by dividing a space into a dominant colour (60%), secondary colour (30%), and accent colour (10%). This project uses image clustering algorithms to quantify how closely any given image adheres to this rule, producing an interpretable compliance score alongside a visualisation of what a "corrected" version of the image would look like.

The exploratory data analysis notebook is available below, with further notebooks to follow.

  • Image clustering
  • Colour spaces
  • Colour theory

Example result

An example image shown alongside the same image after clustering.
An example image alongside its clustered version.
Bar plot of the size of each cluster in the example image.
The size of each cluster in the example image.

Notebooks

Early prototype View Notebook 01 — Exploratory Data Analysis View Notebook 02 — Colour Space & Algorithm Exploration View
Notebook 03 — Scoring Pipeline Coming soon
Notebook 04 — Results Coming soon
Source Code
← Previous Project Next Project →