If a mere apparent color (apparent-color) from an object surface is equal to a color perception as an attribute of the same object surface (surface-color), this condition should be called “perfect color constancy.” If not, the condition should be called “partial color constancy.” We looked into the distinction betweeen these two classes of color constancy, while changing illuminant chromaticity around D65. A small room with variable-chromaticity illuminant was used. The walls were painted with N5 gray. A CRT monitor was placed behind an aperture on the front wall, so that the observer perceived the CRT monitor as a reflecting surface on the wall. Observers were asked to make unique-white settings, in terms of both apparent-color and surface-color, under various chromaticites of illuminants. The shift in unique-white represents the shifts in visual sensitivity, and an analysis based on a simple von Kries type model was made. Cone weights for each cone class were defined to represent the state of chromatic adaptation. The change in cone weights show that the sensitivity shift in the visual system follows illuminant chromaticity change only around white, and the visual system was not able to follow illuminant chromaticity change for saturated illuminants. An additional experiment confirmed that these results had no contamination of simultaneous contrast effect. This result may correspond to the fact that “perfect color constancy” holds under whitish illuminant while “partial color contancy” holds under colored illuminants. In conclusion, we suggest that the “perfect color contancy” can be achieved only by von Kries-like sensitivity change, and the “partial color contancy” requires further processing at the higher stages of the human visual system. key words: perfect color constancy, partial color constancy, chromatic adaptaion, unique-white, cone weight.
ASJC Scopus subject areas
- Electrical and Electronic Engineering