colorClustersKMeans: Cluster pixel colors using K-means clustering

In recolorize: Color-Based Image Segmentation

Cluster pixel colors using K-means clustering

Description

Clusters pixel colors using stats::kmeans().

Usage

colorClustersKMeans(
  pixel_matrix,
  n = 10,
  color_space = "Lab",
  ref_white = "D65"
)

Arguments

pixel_matrix	2D matrix of pixels to classify (rows = pixels, columns = channels).
n	Number of clusters to fit.
color_space	Color space in which to cluster colors, passed to [grDevices]{convertColor}. One of "sRGB", "Lab", "Luv", or "XYZ". Default is "Lab", a perceptually uniform (for humans) color space.
ref_white	Reference white for converting to different color spaces. D65 (the default) corresponds to standard daylight.

Details

Called by colorClusters(). See that documentation for examples.

Value

A list with the following elements:

1. pixel_assignments: A vector of color center assignments for each pixel.

2. centers: A matrix of color centers.

3. sizes: The number of pixels assigned to each cluster.

recolorize documentation built on April 4, 2025, 3:07 a.m.