The relation between color spaces and compositional data analysis demonstrated with magnetic resonance image processing applications

TL;DR

This paper explores how compositional data analysis can be applied to color image processing and MRI data, revealing relationships between color space concepts and compositional components, with applications in image fusion.

Contribution

It introduces a vector decomposition method for compositional data analysis in color and MRI images, linking color space concepts to compositional data analysis for the first time.

Findings

Demonstrated the method on MRI and color images for image fusion

Discussed benefits and disadvantages of the approach in color image processing

Highlighted potential future applications in MRI analysis

Abstract

This paper presents a novel application of compositional data analysis methods in the context of color image processing. A vector decomposition method is proposed to reveal compositional components of any vector with positive components followed by compositional data analysis to demonstrate the relation between color space concepts such as hue and saturation to their compositional counterparts. The proposed methods are applied to a magnetic resonance imaging dataset acquired from a living human brain and a digital color photograph to perform image fusion. Potential future applications in magnetic resonance imaging are mentioned and the benefits/disadvantages of the proposed methods are discussed in terms of color image processing.