Abstract
There is a common belief among perceptual psychologists that shaded images are inherently ambiguous with respect to the sign of relief, and that this ambiguity is resolved by a bias to perceive surfaces illuminated from above. The research described in the present paper was designed to investigate how observers' perceptions of smoothly curved surfaces are influenced by various illumination patterns. The images were generated with a photometrically accurate rendering model that simulates the effects of area lights, cast shadows and indirect illuminations. The displays depicted roughly spherical objects with random patterns of ridges and valleys. These objects could be presented with visible smooth occlusion contours, or these contours could be masked by a uniformly convex aperture. The surfaces could be illuminated from four different directions (i.e., above, below, left or right). Observers judged the pattern of ordinal depth on each object by marking local maxima and minima along designated scan lines. They also judged the apparent magnitudes of relative depth between designated probe points on the surface. There was a high degree of reliability on these tasks both within and between observers. When the different patterns of illumination were compared, the variations in judged depth were remarkably small, and observers' judgments were highly correlated across each possible pair of illumination conditions. They were also highly correlated with the ground truth — even when the occlusion contours were masked. These findings suggest that perceptual ambiguities obtained with unrealistic rendering models may not generalize to more natural contexts.
This research was supported by grants from the National Eye Institute (R01-EY12432) and the National Science Foundation (BCS-0079277).