Investigating accuracy of contour interpolation in modal and amodal completion

Abstract

When information is missing in the visual scene, how does our visual system complete the gap of information? Different methods of perceptual completion have been identified that address different types of gaps, the most apparent types being amodal completion and blind spot filling-in (modal completion); however, the underlying mechanisms of these completion processes are not fully understood. It has been suggested that missing information is completed by interpolating the incomplete contour across the gap. The purpose of this study was to investigate contour interpolation across three different types of gaps (blind spot, occlusion and deletion) by measuring the accuracy and precision of contour representation in each respective type of completion. This was measured with a dot localization paradigm which required interpolating an incomplete contour across the gap of information. Performance data in each condition was collected and compared. Overall, we found that the contours interpolated behind an occluder (amodal completion) were the most inaccurate and imprecise while contour interpolation across the physiological blind spot by blind spot filling-in (modal completion) had similar accuracy and precision to the contours interpolated across a simple gap where information was absent (deletion). We also found that all three types of perceptual completion interpolated the curvilinear contour stimulus flatter than their theoretical paths, of which, blind spot filling-in produced the flattest completions. Our study confirms previous findings regarding the shape of interpolated contours and also extends the understanding about the precision and accuracy of blind spot filling-in and modal completion, as well as amodal completion.

Keywords: perceptual completion, blind spot filling-in, dot localization paradigm, contour interpolation, deleted stimulus