Volume Completion

Abstract

The visual system completes image fragments into larger regions when those fragments are taken to be the visible portions of an occluded object. Kellman and Shipley (1991) argued that this "amodal" completion is based on the way that the contours of image fragments "relate." Contours relate when their imaginary extensions intersect at an obtuse or right angle. However, it is shown here that contour relatability is neither necessary nor sufficient for completion to take place. Demonstrations that go beyond traditional examples of overlapping flat surfaces reveal that "mergeable" volumes, rather than relatable contours, are the critical elements in completion phenomena. A volume is defined as a 3-D enclosure. Typically, this refers to a surface plus the inside that it encloses. Two volumes are mergeable when their unbounded visible surfaces are relatable or the insides enclosed by those surfaces can completely merge. Two surfaces are relatable when their visible portions can be extended into occluded space along the trajectories defined by their respective curvatures so that they merge into a common surface. A volume-based account of amodal completion subsumes surface completion as a special case and explains examples that neither a contour- nor a surface-based account can explain.

Curvature discontinuities are cues for rapid shape analysis

“…Although the sharpness of curvature discontinuities can raise several ecologically plausible interpretations, in isolation they cannot disambiguate between them. To accomplish this, certain assumptions about ground plane orientations, viewing angles, and the most probable shapes of objects in the world may be required (Albert & Tse, 2000;Tse, 1998Tse, , 1999aTse, , 1999bTse, , 2000.…”

The “side” matters: How configurality is reflected in completion

“…If the neurons active at the illusory contours in the Kanizsa figure are indeed BOWN-sensitive neurons, it would make a strong case against the "line-drawing" point of view because (1) BOWN signals do not merely indicate the existence of a depth difference, but also indicate the polarity of the difference and (2) BOWN signals reflect the global configuration of an image. It would also make a strong case for the involvement of depth-order computation in creating illusory contours (Anderson & Julesz, 1995;Coren, 1972;He & Nakayama, 1992;Nakayama et al, 1995;Nakayama & Shimojo, 1990Nakayama et al, 1989;Tse, 1999aTse, , 1999b.…”

“…In fact, some papers showed that the perception of depth order or 3-D configuration plays the fundamental role in the completion phenomenon (Anderson & Julesz, 1995;He & Nakayama, 1992;Nakayama, He, & Shimojo, 1995;Nakayama & Shimojo, 1990Nakayama et al, 1989;Tse, 1999aTse, , 1999b. Importantly, the determination of depth order is highly context-sensitive.…”

Amodal volume completion: 3D visual completion

“…In this paper, we review prior psychological theory and computer vision to identify the supporting aspects for this view of completion with specific reference to the completion of unknown 3D volumes-considered here as the completion of the exterior bounding surface(s) for an unknown (i.e., unbounded) enclosure within a given 3D scene. Notably, we see that 3D completion through visual propagation is specifically supported by the recent psychological account of Volume Completion [72], itself surpassing earlier work in this field, although computer vision work in considering such aspects remains limited.…”