To examine the effect of perspective information on stereo-depth aftereffects, we compared the magnitudes of the aftereffects under two conditions: the cue-concordant condition, in which binocular disparity and perspective specified the slant of adaptation stimuli in concordance, and the cue-conflict condition, in which binocular disparity specified the slant of the stimuli, while perspective specified the fronto-parallel plane. We found that the subjects who relied predominantly on the disparity-specified depth in slant estimation showed larger aftereffects in the cue-conflict condition, suggesting that the recalibration of the relationship between disparity and perceived depth facilitated by the cue-conflict is responsible for the aftereffects. However, the subjects who relied predominantly on the perspective-specified depth in slant estimation showed larger aftereffects in the cue-concordant condition, suggesting that in these subjects the aftereffects were mainly mediated by neural fatigue. Our results suggest that the stereo-depth aftereffects were caused by multiple factors such as fatigue and recalibration and their relative contributions depend on the cue weighting.

To reconstruct shape from shading, the illuminant direction is important. Many researchers have reported that human visual systems extract the illuminant direction from visual information for perceiving shape from shading. Pentland (1982) proposed that only the luminance distribution was needed to extract the illuminant direction from visual information. Besides, Berbaum et al. (1984) proposed that shape perception of other objects along with the surface luminance distributions was needed. However, Berbaum et al. (1984) pointed out that these two models haven’t been separately investigated in previous researches. In this research, we investigated how the human visual system extracts the illuminant direction from visual information for perceiving shape from shading to test the validities of the models. For the purpose, we manipulated independently binocular disparity to specify shape perception and luminance distribution of the stimulus. The results implied that both Pentland and Berbaum model were valid and that human visual system extracts the illuminant directions from the two systems and chose one of them depending on the stimulus situation for perceiving shape from shading, although there were individual differences.