Color space of normally sighted and color-deficient observers reconstructed from color naming

An experimental procedure based on the color-naming method introduced by Boynton, Schafer, and Neun (1964) was used to study the color appearance of equiluminant spectral stimuli in observers with congenital red-green color deficiencies, as well as in normal trichromats. subjects' responses (choice of one or more labels from the set red, yellow, green, blue, and white) were converted to numeric scores, which were used to estimate subjective differences between pairs of colors. Individual subjects' matrices were processed by means of multidimensional scaling. As in the direct rating of color dissimilarities in normal trichromats (Sokolov & Izmailov, 1983) and color-deficient observers (Paramei, Izmailov, & Sokolov, 1991), these indirectly obtained measures yielded a color space in which three dimensions appear to be necessary and sufficient. The dimensions are interpreted as evidence for red-green, blue-yellow, and achromatic (saturation) subsystems. Based on the color-naming technique, three-dimensional spaces were reconstructed for the color-deficient observers. These results were compared with those obtained by Helm (1964). It is argued that retaining more than one (blue-yellow) dimension in the color spaces of such observers provides additional information indicating preservation of residual red-green discrimination accompanied by finer discrimination of chroma than in normal trichromats. The spherical model of color discrimination developed for normal trichromats (Izmailov & Sokolov, 1991) is shown to be valid for color-deficient subjects as well and may be useful as a framework for differentiating protan and deutan types of color deficiency. color-naming functions, which seem not to reveal a differentiation between protans and deutans, provide results from which this differentiation can be extracted in reconstructed color spaces.