Abstract

Robot-vision models of color constancy that are based on a linear approximation of illuminant and reflectance spectra are often generalized to the human visual system. Dannemiller's¹ computational approach to color constancy is a good example. In this paper it is shown that such a procedure to estimate illuminant and reflectance spectra is less plausible for the human visual system than is implicitly assumed in robot-vision. The resemblance of the modeled hypothetical visual systems to the human visual system is misleading since it implies that surface reflectance is the illuminant-invariant object color descriptor the human visual system uses to achieve color constancy. However, an alternative type of descriptor is available that is not used to recover reflectance spectra. It has the advantage of allowing an interpretation that is preferable from the human point of view.

© 1990 Optical Society of America