Abstract
Large background surfaces serve as a reference frame for egocentric target localization in the intermediate distance range. The visual representation of these surfaces (creating our perceptual space) relies on the visual system’s intrinsic bias and external depth information. The intrinsic bias below the eye level is an implicit slant surface, and is revealed by measuring perceived locations of dimly-lit targets in the dark (Ooi et al, 2001). To understand how the terrestrial environment shapes visual processing, we explored whether space perception below the eye level is more accurate than above the eye level. Experiment 1 obtained the intrinsic bias using a blind walking-gesturing task that measured the perceived location of a dimly-lit target (0.2 deg) at 5 distances (1.5-7 m) and 2 heights (above and below the eye level) in the dark. We found the mean results for both the above and below eye level conditions exhibit the familiar intrinsic bias. However, the perceived distance is moderately farther (less underestimation) in the below eye level condition, suggesting some ground advantage. Experiment 2 tested the possibility of a ground advantage with weak texture cues (two parallel rows of dimly-lit elements) in a reduced cue environment. We measured the perceived target location in the presence of the texture array that was pasted 1.25m above, or below, the eye level. We found that although perceived distances are underestimated in both ground and ceiling texture conditions, the underestimation is significantly smaller in the former. This ground advantage, which is stronger than in Experiment 1, is also found when observers verbally reported the egocentric target distance. Altogether, our findings demonstrate that background surfaces below the eye level (ground-like), are more efficiently represented. The enhanced perception likely reflects an adaptation to our terrestrial existence (Bian et al, 2005; Gibson, 1950; McCarley & He, 1999; Sedgwick, 1986).
Meeting abstract presented at VSS 2012