Purchase this article with an account.
Marianne Maertens, Robert Shapley; Local determinants of contour interpolation. Journal of Vision 2007;7(9):110. doi: https://doi.org/10.1167/7.9.110.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Object parts in the natural environment are often excluded from view due to occlusion. The visual system adapted to such degraded viewing conditions by developing a remarkable ability to interpolate linking contours between isolated object fragments. One factor that is believed to be of critical importance for the perceptual strength of interpolated (or “illusory”) contours is the Support Ratio, the ratio between luminance-defined contour support and total contour length (Kellman & Shipley, 1991).
In Varin figures in which vivid illusory contours are interpolated between inducers (Lesher & Mingolla, 1993), solid Kanizsa inducers are replaced by a number of partially-occluded concentric arcs. Their luminance-defined contour support is dramatically reduced (compared with corresponding Kanizsa-type inducers) because the line ends amount to only a small proportion of the total contour length. Here we tested whether the crispness of the perceived illusory contour in Varin figures is a function of the luminance-defined contour support, or alternatively, depends on the distance between the line-endings of the inducers. We independently manipulated the line density and line width of the inducer circles, resulting in conditions with equal Support Ratio but different line density, and vice versa. We measured subject's performance in a probe localization task (Stanley & Rubin, 2001; Guttman & Kellman, 2004), where a small target is presented at different distances from a curvilinear illusory contour (Ringach & Shapley, 1996) and subjects have to indicate whether the target appeared inside or outside the illusory figure. Sensitivity, as indicated by the slopes of the psychometric functions for the accuracy of probe localization, increased with increasing line density -independent of Support Ratio. This is objective evidence that the perceived crispness of illusory contours in the Varin figures was determined by the separation between 2-d features, the inducers’ line endings, and not by Support Ratio.
This PDF is available to Subscribers Only