Abstract
To accomplish color constancy the illuminant color needs to be discounted from the light reflected from surfaces. Some strategies for discounting the illuminant color use statistics of luminance and chromaticity distribution in natural scenes. It was shown that the luminance-chromaticity correlation of surfaces was a strong cue for estimating the illuminant color (Golz and MacLeod, 2002). In this study we show in what degree color constancy exploits the potential cue that is provided by the luminance balance of differently colored surfaces. In our experiments we used six colors: bright and dim red, green and blue, as surrounding stimulus colors. The center stimulus served as the test stimulus, and its chromaticity was adjusted by the observer to obtain a perceptually achromatic surface. Three simulated black body radiations of 3000, 6500 and 20000K were used as test illuminants. The luminance and chromaticity of surrounding colors were independently controlled to see each effect on color constancy. In some conditions bright surrounding colors were set to be optimal colors under the corresponding illuminant. The results showed that the chromaticity shifts of surfaces associated with an illuminant change had larger effects than changes in relative luminance only. But changes in the luminance balance of surfaces (with no chromaticity shift) also showed clear effects on the estimate of illuminant color. Furthermore highly saturated and brighter colors had greater influence than low saturated and dim colors, consistent with a statistically motivated model of illuminant estimation.