Color representations of normals and congenital red–green color deficiencies: Estimation of individual results based on color vision model. Issue 3 (13th December 2021)

• Record Type: Journal Article
• Authors: Ohkoba, Minoru; Ishikawa, Tomoharu; Hira, Shoko; Ohtsuka, Sakuichi; Ayama, Miyoshi
• Is Part Of: Color research & application. Volume 47:Issue 3(2022)
• Page Start: 565
• Page End: 584
• Publication Date: 2021-12-13
• Subjects: congenital color vision deficiency -- deutan -- hue circle -- MDS -- protan; Color -- Periodicals; 535
• DOI: 10.1002/col.22763
• Languages: English
• ISSNs: 0361-2317

Abstract

Abstract: Difference scaling experiment was conducted to investigate the mental color representation of congenital color‐deficient observers and observers with normal color vision. Two decks of cards, high and medium chroma, each containing 10 Munsell chips, were prepared. A total of 45 pairs of hues were prepared for each of the decks. Ten protans, 10 deutans, and 10 people with normal color vision participated as observers, with each asked to rate the perceptual distance between two colors on a given color card. The results were analyzed using the multidimensional scaling method. All observers with normal color vision showed a circular shape close to the Munsell hue circle, while the majority of color‐deficient observers showed a concave shape bending at Y and PB. To indicate the degree of distortion from a circle, the distortion index was proposed to quantitatively evaluate intergroup and individual differences. To investigate the underlying mechanism of intergroup differences as well as individual differences in color representations of observers with normal color vision and color‐deficient observers, we proposed a model that considers various levels of human color vision mechanism from the cone pigment absorption, the luminance and opponent‐color coding level, and nonlinear transformation to difference‐scaling judgment. The circular shape for observers with normal color vision and some color‐deficient observers, as well as concave shapes for most color‐deficient observers were estimated. The correlation coefficient between the estimation and experiment‐based difference ranged from r = 0.64 to r = 0.94 with the grand average of r = 0.82, with p ‐values less than 0.001 for all observers, suggesting that the concept of proposed model is appropriate. Abstract :