Abstract
The anchoring theory of lightness perception proposes that the perception of white is linked to the brightest surface in a scene. A lot of evidence for some form of anchoring has been amassed, but mainly under photopic illumination conditions. There is hardly any research on lightness perception under scotopic conditions, when only rod photoreceptors are active. We tested whether anchoring also works under these conditions. In other words: are all cats really gray at night, even white ones? We printed 10 chips, such that they were equally spaced on the L* scale of the CIELUV color-space (under a D65 illuminant). 6 naive observers first viewed the maximally and minimally reflecting chips at 277 cd/m^2. They were instructed to memorize their percepts as 100% white and 0% white. Observers adapted to three light levels (1.2x10^(-4) (scotopic), 28, and 277 cd/m^2; counter-balanced) and viewed our chips, one at a time, in two, randomly-ordered blocks. For scotopic adaptation, observers wore fitted goggles with neutral density filters. They adapted for at least ten minutes and dark adaptation was confirmed by a foveal scotoma and the inability to sort colored chips by hue. Observers reported the perceived amount of white in each chip, in units of 10% along their memorized white scale. At both photopic levels, observers assigned an average rating of 95.85% to the maximally reflecting chip. When dark adapted, observers assigned an average rating of 75.83% to the maximally reflecting chip, with a monotonic decline for darker chips. This agrees well with our own phenomenological observation that the white chip appears gray in the scotopic range. We propose that cone activation may be necessary to perceive white.
Meeting abstract presented at VSS 2015