Abstract
Purpose: Prior theories of the perception of partially occluded figures have stressed the role of contour information. Meanwhile, the role of region information has been less understood. We investigated the extent and nature of the contribution of region to the perception of partially occluded figures. Method: Each stimulus contained a black figure against white ground. Each figure was composed of 4 rectangles, so that contours of a figure had either horizontal or vertical orientation. In Experiment 1, white diamond distractors were used as occluding elements, while in Experiment 2, white square distractors were used as occluders. In both experiments, 2 factors were manipulated: presence of region information (present vs. absent) and occluder size (small vs. large). Thus, each experiment had 4 conditions. Stimuli in the ‘region absent’ conditions were created from those in the corresponding ‘region present’ conditions by displaying only edges (contours). On each trial, a subject's task was to respond whether the occluded figure was displayed in its upright or upside-down orientation. Results: In both experiments, performance in the ‘region present’ conditions was superior to that in the corresponding ‘region absent’ conditions. This was true whether occluders were large or small, although the effect of region was greater when small occluders were used.
Conclusions: These results suggest that contours alone cannot account for perception of partially occluded figures. Also, because white diamond occluders were used in Experiment 1, it might be argued that these occluders gave rise to illusory diamonds appearing ‘in front’ of the figure. Thus, depth information would have been available in the ‘region present’ conditions of Experiment 1. However, the results of Experiment 2 show that depth cues cannot wholly explain the effect of region — white square occluders did not give rise to illusory squares appearing ‘in front’ of the figure in ‘region present’ conditions.