Illusory Occluding Contours and Surface Formation by Depth Propagation

Abstract

A novel kind of depth-spreading effect which should be distinguished in various aspects from the known interpolation, averaging, or 'filling-in' phenomena is reported. The demonstrations and experiments suggest that depth from an uncrossed disparity can be extrapolated from, not just interpolated between, illusory or real contours to form perceptually a background surface. In addition, the form of the illusory contour itself could be drastically changed in configuration and sharpness, contingently with perceptual background-surface formation. No such effects of surface and contour formation were observed in the crossed disparity case. Because the illusory contours were enhanced and perceived as illusory 'occluding contours', these effects may be closely related to the 'occlusion constraints' in the real world.

Improving relative depth judgments in augmented reality with auxiliary augmentations

“…Additionally, in front of the wall, the virtual window is not captured in depth, and thus only the AOI is localized in the near depth, making the depth judgment easier. In this study, a rendered virtual window forms a "physical-like" surface, which is not an illusory contour as in the study of Takeichi et al [1992], for which an illusory contour was created with Kanizsa figures [Kanizsa 1979]. The use of illusory contours in AR visualizations has not been investigated, but they can be useful for making virtual windows and surfaces because they require less screen space than real contours.…”

“…In addition to depth cue theories, stereoscopic depth perception is affected by local depth interactions between surfaces. Examples of these include a depth interpolation and extrapolation [Collett 1985], depth capturing [Ramachandran and Cavanagh 1985], depth propagation [Takeichi et al 1992], and the Da Vinci phenomenon [Nakayama 1996]. The depth cues provide higher-level depth information for human visual processing, whereas these phenomena affect lower-level vision.…”

3-D vision and figure-ground separation by visual cortex

“…These effects may be explained by a filling-in process that selectively completes a BC surface representation at a depth corresponding to that of region CD. A variety of recent experiments have demonstrated that a filling-in process does, indeed, complete various depthful surface properties (Nakayama, Shimojo, & Ramachandran, 1990;Nakayama, Shimojo, & Silverman, 1989;Takeichi, Watanabe, & Shimojo, 1992;Watanabe & Cavanagh, 1992). To explain how this occurs, the theory utilizes the following types of processes.…”

“…These include studies of texture segregation (Beck, Graham, & Sutter, 1991;Beck, Rosenfeld, & Ivry, 1990;Cruthirds, Grossberg, & Mingolla, 1993;Graham, Beck, & Sutter, 1992;Sutter, Beck, & Graham, 1989), border effects on color detection (Eskew, 1989;Eskew, Stromeyer, Picotte, & Kronauer, 1991), visual phantoms (Brown & Weisstein, 1988a), 3-D surface formation from 2-D textures (Buckleyet al, 1989;Todd & Akerstrom, 1987), interactions between filling-in of brightness or color and illusory contour formation (Dresp, Lorenceau, & Bonnet, 1990;Field, Hayes, & Hess, 1993;Kellman & Shipley, 1991;Lesher & Mingolla, 1993;Nakayama, Shimojo, & Ramachandran, 1989;Prinzmetal, 1990;Prinzmetal & Boaz, 1989;Ramachandran, 1992;Shipley & Kellman, 1992;Takeichi, Shimojo, & Watanabe, in press;Watanabe & Sato, 1989;Watanabe & Takeichi, 1990), interactions between depth, emergent segmentation, and filling-in (Meyer & Dougherty, 1987;Nakayama, Shimojo, & Ramachandran, 1990;Takeichi, Watanabe, & Shimojo, 1992;Watanabe & Cavanagh, 1992), orientation-specific luminance aftereffects (Mikaelian, Linton, & Phillips, 1990), transient dynamics of filling-in (Arrington, 1992(Arrington, , 1993Paradiso & Nakayama, 1991), cortical dynamics of emergent segmentation (Peterhans & von der Heydt, 1989;von der Heydt,...…”

Depth capture by generic‐view motion

“…It is known that unambiguous cues, such as disparity, occluding contour, or motion with salient features, lead to the capture of an ambiguous presentation of other elements/regions Depth capture by generic-view motion 87 (e.g., Mather, 1989;Mitchison & McKee, 1987a, 1987bRamachandran & Cavanagh, 1985, 1987Takeichi, Watanabe, & Shimojo, 1992). In the present study, the visual element to be captured was the rotating cube, the presentation of which was ambiguous, especially when the image offered was an accidental view, while the unambiguous cues were either random dots (Experiment 1) or bars (Experiment 2).…”