Abstract
Although amodal completion is known to occur under conditions of occlusion it is not known whether perceived occlusion in interrupted objects is determined by the same factors as good continuation. We measured the effects on perceived occlusion of gap size and contrast polarity, both of which influence good continuation. Stimuli were sets of 5 rectangles at varying near vertical orientations, also varying in length, width and separation and interrupted by a horizontal gap. The degree to which the horizontal gap was perceived as an occluding surface was compared across stimuli using the method of paired comparison. In Experiment 1 we used two gap sizes (17′ and 52′) and three luminance arrangements on a grey background. Stimuli were either all black, alternating black and white with the same contrast polarity across the gap, or alternating with opposite contrast polarity. Multiple randomly generated versions of each stimulus were created. For all stimuli, perceived occlusion was significantly stronger for the large gap than the small gap. Change in luminance polarity across the gap generally reduced the sense of occlusion but significantly more for the small gap. Experiment 2 tested the possibility that change of contrast (rather than polarity) reduced the sense of occlusion in Experiment 1. Using only the small gap, we compared opposite polarity with same polarity conditions controlling for contrast differences between components (both relative to each other and relative to the background). Contrast change across the gap did impair perceived occlusion but significantly less than polarity change. We conclude that unlike amodal completion, the sense of occlusion at gaps in continuing objects is greater for large gaps than for small gaps. The considerably stronger effect of polarity change on perception of small gap occlusion may indicate a greater role of good continuation in that case than for large gap occlusion.
ARC Grant RM02837 to B.G. and B.A.