Contour Integration Across Polarities and Spatial Gaps: From Local Contrast Filtering to Global Grouping

Abstract

This article introduces an experimental paradigm to selectively probe the multiple levels of visual processing that influence the formation of object contours, perceptual boundaries, and illusory contours. The experiments test the assumption that, to integrate contour information across space and contrast sign, a spatially short-range filtering process that is sensitive to contrast polarity inputs to a spatially long-range grouping process that pools signals from opposite contrast polarities. The stimuli consisted of thin subthreshold lines, flashed upon gaps between collinear inducers which potentially enable the formation of illusory contours. The subthreshold lines were composed of one or more segments with opposite contrast polarities. The polarity nearest to the inducers was varied to differentially excite the short-range filtering process. The experimental results are consistent with neurophysiological evidence for cortical mechanisms of contour processing and with the Boundary Contour System model, which identifies the short-range filtering process with cortical simple cells, and the long-range grouping process with cortical bipole cells.

Dynamic characteristics of spatial mechanisms coding contour structures

“…While Yu and Levi's (1997a) data show spatial facilitation with short targets and inducers within spatial limits up to 20 arcmin only, other findings with much longer targets and inducing lines indicate that spatial facilitation occurs within larger spatial scales as well (Dresp and Grossberg, 1997;Wehrhahn and Dresp, 1998). Here again, the conclusions are not straightforward.…”

“…This surprising result may reflect the output of a mechanism of contour grouping that is activated only when the decreasing proximity of stimulus fragments is compensated by stronger local contrasts, whichever sign they carry. Dresp and Grossberg (1997) conducted spatial facilitation studies with edge-like inducers and fractioned line targets collinear with the orientation of the inducing edges. In some conditions, the contrast sign of half the target line was the same as that of the nearest inducer, in others it carried the opposite sign.…”

“…In some of the studies with line targets and collinear inducers (e.g. Dresp and Grossberg, 1997), it has been necessary to run subjects in several thousands of training trials with the inducing contours before the lowest possible detection threshold was obtained. Interestingly, in the control conditions where the line target is presented without contour inducers, detection performances evolve only very little, if at all.…”

“…Psychophysical experiments on detection facilitation with collinear targets and inducers, now called .spatial facilitation (Yu and Levi, 1997a), have generated a coherent body of data supporting hypotheses about visual mechanisms of spatial grouping and the early processing of object contours and illusory contours Bonnet, 1991, 1993;Dresp, 1993;Sagi, 1993, 1994a, b;Dresp and Bonnet, 1995;Kapadia et al, 1995;Dresp and Grossberg, 1997;Yu and Levi, 1997a;Wehrhahn and Dresp, 1998).…”

The influence of contrast and spatial factors in the perceived shape of boundaries

“…Studies on spatial facilitation have demonstrated that high-contrast-inducing lines facilitate the detection of a target line of either contrast polarity when the distance separating them is wider than 20 arcmin (Wehrhahn & Dresp, 1998). Experiments with edge-like inducers and collinear line targets demonstrated detection facilitation with shorter spatial separations, but only in conditions in which the inducers and the target had the same contrast polarity (Dresp & Grossberg, 1997). Clearly, the role of contrast polarity in contour integration is still an open question.…”