Abstract
The completion of partly occluded objects appears instantaneous and effortless. However, research indicates that interpolation takes measurable time, and that time-to-completion depends on a number of stimulus variables. The current study asks how the time course of visual completion depends on the amount of occlusion, and examines the mechanisms underlying temporal variations. Experiment 1 used a primed-matching paradigm to determine completion times for circles and squares occluded by different amounts: 20% and 32.5% contour occlusion. Experiments 2 and 3 used a dot localization paradigm to probe completed contour representations for a qualitative shift as the amount of occlusion exceeds some spatial limit. Our results provide the first demonstration that, if given sufficient processing time, highly occluded objects — including some with visible contours exceeding the limits of spatial relatability — achieve functional equivalence to their complete counterparts. This finding suggests that the visual system can complete highly occluded objects, although time-to-completion rises with amount of occlusion. Furthermore, the dot localization results indicate that the precision and shape of the interpolated contour representations vary smoothly with amount of occlusion, even though the visible edges in the more highly occluded shapes were non-relatable (because the interpolated edge needed to bend through more than 90 degrees). Thus, increases in completion times do not appear to result from a breakdown of low-level interpolation processes beyond some spatial limit: The same contour completion mechanism operates on objects occluded by different spatial extents. Implications for models of boundary interpolation will be discussed.