PUPMCR: an R package for image-based identification of color based on Rayner's (1970) terminology and known fungal pigments.

Abstract

Fungi are eukaryotic organisms grouped based on different traits of their morphology. In 1970, R. W. Rayner published A Mycological Colour Chart to provide a standardized system for identifying color in fungi. While its terminologies have contributed a standard way of color matching for taxonomic diagnoses, this method using the personal color perception of the observer does not guarantee accuracy. Considering the diversity of fungi, visual color matching is expected to be challenging without a standard assisting instrument. In this study, the R package PUPMCR is developed to approximate the color name and associated pigments of fungal species based on the pixel coordinates of its uploaded image. This software utilizes CIELAB and RGB color spaces as well as Euclidean and Chi-square distance metric systems. The package is tested and validated using 300 fungal images as a dataset for conducting interrater reliability tests. Results showed the highest agreement for parameters utilizing the RGB color space (Cohen's kappa values: 0.655 ± 0.013 for RGB and Euclidean; 0.658 ± 0.004 for RGB and Chi-square), attributed to its computational efficiency, which facilitates more uniform binning and universally scaled distance metrics. The produced color-identifying tool is also available as a Shiny web application (https://pupmcr.shinyapps.io/PUPMCR/) to allow better accessibility for users on the World Wide Web. The development of PUPMCR not only benefits a variety of users from its free accessibility but also provides a more reliable color identification system in the field of mycology.