Abstract
Purpose Ecologically, T-junctions and L-junctions in the image are concurrent with situations of occlusion: they arise when the bounding contour of the occluding surface intersects with that of the occluded surface. While junctions were previously used for visual processing in computational studies, their role in human perception has not been demonstrated directly. Methods We constructed stimuli that differ only in the junction structure, keeping other variables the same. 1, In the modal completion ("illusory contours") case, the L-junctions on the inducing elements Cpacmen') of a Kanizsa square were covered by small opaque circles. Subjects were asked to judge the strength of the illusory contours. 2, In the amodal completion case, we used a portion of an occluded surface that was visible between two occluders as a test-patch, and asked subjects whether the patch appeared as part of the occluded surface or not when the T-junctions were experimentally altered (see example below, top figure; control condition, bottom). We used stereo and motion to enhance the effects and facilitate subjects' decision; however, this is not a necessary condition and the results can be observed in static, monocular images. Results Subjects reported not perceiving illusory contours at all, and perceiving the test patch as being in front, when the L-junctions and T-junctions, respectively, were manipulated as described. In the motion case, the T-junction structure determined the perceived motion of (he test patch. Conclusions The perception of occlusion depends critically on the presence of I junctions in the image, even when other occlusion cues (relatability, high-level interpretation) are left unchanged.
Original language | English (US) |
---|---|
Pages (from-to) | S203 |
Journal | Investigative Ophthalmology and Visual Science |
Volume | 38 |
Issue number | 4 |
State | Published - 1997 |
ASJC Scopus subject areas
- Ophthalmology
- Sensory Systems
- Cellular and Molecular Neuroscience