nearest-human-color

Find the nearest approximation of a color given a predefined list of colors.

Usage

var colors = {
  red: '#f00',
  yellow: '#ff0',
  blue: '#00f'
};

var nearestHumanColor = require('nearest-human-color').from(colors);

nearestHumanColor('#800'); // => { name: 'red', value: '#f00', rgb: { r: 255, g: 0, b: 0 }, distance: 119 }
nearestHumanColor('#ffe'); // => { name: 'yellow', value: '#ff0', rgb: { r: 255, g: 255, b: 0 }, distance: 238 }

How it works

Finding the nearest color is a specific case of the "nearest neighbor search" (or NNS) problem. The predefined colors can be thought of as points in 3D space where the X, Y, and Z axes represent each color's red, green, and blue (RGB) values. So finding the nearest color to any given value amounts to finding the closet neighbor to the point where that color would reside when plotted in such a 3D space.

From the Wikipedia article on the subject:

The simplest solution to the NNS problem is to compute the distance from the query point to every other point in the database, keeping track of the "best so far". This algorithm, sometimes referred to as the naive approach, has a running time of O(Nd) where N is the cardinality of S and d is the dimensionality of M. There are no search data structures to maintain, so linear search has no space complexity beyond the storage of the database. Naive search can, on average, outperform space partitioning approaches on higher dimensional spaces.

This library uses an approach to approximate what the human eye sees. Performance should be totally fine unless there are many pre-defined colors to search (and even then, it will probably only matter if you're calling nearestHumanColor a ton of times).

The library then takes the results from the Euclidean color differencial, and then average it with CIEDE2000 color differencial calculation to obtain an approximation of closest color as a human eye would see.

The approach is a combination of the two methods listed here, along with a factor for hue as well, with some tweaking to imrpove output: https://en.wikipedia.org/wiki/Color_difference

The most realistic optimization that could be made here would probably be to cache results so that multiple calls for the same color can return immediately.

Special Thanks

Special thanks to the following individuals as portions of their projects were used for this project.

dtao The code in this project was originally forked from their nearest-color library and would not be possible without this impressive work. https://github.com/dtao/nearest-color
Qix- The code used for the rgb to lab conversion is from this project. https://github.com/Qix-/color-convert
markusn The CIEDE2000 color differencial algorithym is from this project. https://github.com/markusn/color-diff

peter-lewis / nearest-human-color Goto Github PK