Abstract
We asked whether the spatial expanse of scene views can be judged across changes in object orientation based on predicted differences in boundary extension (BE) as a function of object orientation. Stimuli were photographs of overhead scene views showing an object against a natural ground surface (e.g., a flashlight on grass), shown within a square view-window. Objects filled views in one dimension only (e.g., a screwdriver, hammer). On each trial, observers (N=48) viewed a 250-ms picture followed by a 250-ms mask and test picture. Half the objects were vertical; the other half horizontal. On half the trials, the object in the test view was in the same orientation as the stimulus view (VV and HH trials); in the other half, it was in the complementary orientation (VH and HV trials). Stimulus and test views were always close-ups. Horizontal-vertical viewing information was made comparable by placing square frames in front of monitors, with square cutouts for stimuli. At test, observers rated the test picture’s view as the same view, more close-up, or more wide-angle than the stimulus view on a continuous sliding scale. A negative rating significantly less than zero ("same view") would indicate BE. We hypothesized that vertical-object stimuli would elicit more BE than horizontal-object stimuli (as in Dickinson et al., VSS 2011), resulting in VH stimulus-test pairs being rated more similar than HV pairs. In addition to finding significant BE for VV and HH trials (mean ratings: –.19 and –.13; both ps <.001), we found a VH-HV asymmetry: VH pairs were rated as more similar than HV pairs (–.05 vs. –.28; p <.001), which is analogous to the CW-WC asymmetry indicative of BE. The results suggest that observers can judge the spatial expanse of views independent of object orientation and provides further support for the finding that vertical-object scene views are perceived as more close-up than horizontal-object scene views.
Meeting abstract presented at VSS 2012