Abstract
The visual appearance of real world objects depends on their reflectance properties, illumination and geometrical factors. We asked how human colour constancy performance deals with the depth-mesostructure of surfaces, which may provide additional cues or, alternatively, worsen colour constancy. In our experiments we presented samples (circular patches, 90 mm) of different materials (paper, fur, tile, cloth), each mounted on a black sample holder (front 30 x 15 cm, 60 deg slope) and presented in the middle of a black viewing box (1.0 x 1.0 x 0.8 m). All samples appeared nearly achromatic under daylight. The samples and the box were illuminated by a computer controlled, calibrated LCD-projector (Panasonic PT AE 1000E) which was mounted above the observers’ head. The observers (2 male, 2 female) viewed the samples frontally (viewing distance 90 cm), either with their head fixed or while making sideway movements with their upper body. The chromaticity of the standard illuminant was D65, the equiluminant test-illuminants were chosen from one of the cardinal axes. After viewing the samples under the standard illuminant for 15 s, the respective test-illuminant was presented for 5 s. Colour constancy was quantified by an achromatic setting method, whereby the degree of colour constancy was calculated from the shift of the achromatic colour locus associated with the illumination change (for illuminant D65, the achromatic colour locus of all samples was within 1 Delta EUV of the colour locus of D65). In the fixed-head condition, we found no differences in the amount of colour constancy achieved for the different materials. However, colour constancy for the sample with the most depth-mesostructure (fur) was significantly increased by selfmotion of the observer. We discuss the possible exploitation of cues from selfmotion and the mesostructure of the material.
Meeting abstract presented at VSS 2012