colordistance

colordistance quantifies color distributions in digital images by creating multidimensional color histograms and computing pairwise Earth Mover's Distance (EMD) to enable objective comparison of color profiles in biological research.


Key Features:

  • Quantitative Color Profiling: Processes digital images to quantify color profiles, extract trait data, and compare color palettes by similarity and abundance.
  • Multidimensional Color Histograms: Represents image pixels as 3D coordinates within a specified color space and generates multidimensional histograms capturing color distribution.
  • Earth Mover's Distance (EMD): Computes pairwise distances between color histograms using EMD to measure transportation-cost-based color similarity and difference.
  • Customizable Parameters: Allows selection of color spaces and adjustment of histogram generation and pairwise comparison parameters to produce tailored color distance matrices.

Scientific Applications:

  • Species Identification: Enables objective comparison of coloration for species discrimination and taxonomic studies.
  • Phenotypic and Seasonal Variation: Quantifies breeding- or season-related changes in coloration and other phenotypic color traits.
  • Behavioral and Ecological Coloration Studies: Supports analysis of coloration relevant to defense, camouflage, and other ecological or adaptive roles.
  • Image-based Trait Quantification: Provides a quantitative alternative to subjective digital image classifications when spectrophotometry or specialized equipment are impractical.

Methodology:

Images are processed to extract pixel data as coordinates within a chosen color space; multidimensional color histograms are generated for each image; pairwise distances between histograms are computed using Earth Mover's Distance (EMD) to produce color distance matrices.

Topics

Details

License:
GPL-3.0
Maturity:
Mature
Cost:
Free of charge
Tool Type:
library
Operating Systems:
Linux, Windows, Mac
Programming Languages:
R
Added:
5/19/2019
Last Updated:
6/16/2020

Operations

Publications

Weller HI, Westneat MW. Quantitative color profiling of digital images with earth mover’s distance using the R package colordistance. PeerJ. 2019;7:e6398. doi:10.7717/peerj.6398. PMID:30775177. PMCID:PMC6371918.

PMID: 30775177
PMCID: PMC6371918
Funding: - NSF: IOS 1425049 and DEB 1541547

Documentation