No abstract

The possibility that subjective contours are an artifact of brightness contrast was explored. In one experiment, inducing luminance was found to have different effects on the clarity of subjective contours and the magnitude of brightness contrast. The results of a second experiment indicated that differences of luminance in a stimulus display are necessary for subjective contours to be sustained, whereas chromatic differences are not. It is concluded that subjective contour and brightness contrast are distinct perceptual phenomena but share a dependency on the processing of edge information transmitted through the achromatic channels of the visual system.

This experiment demonstrates what we call a proximity effect in forced-choice v~ua! dete~tion: Detection accuracy improves as the distance is increased between the target and the nOise Items III tile array that are confusable with it. The proximity effect is a natural prediction of Este~'s theory t~at detection is mediated by feature-detecting receptive fields, but other recent models of vISual detection do not predict it. However, the pattern o~ results see.ms best explain~d ~n t~r~~. of 'perceptual configurations in the array; when the target IS grouped WIth confusable nOIse ItS vlslblhty IS less than when it is not.In the forced-choice visual detection paradigm, the S is given brief presentations of arrays of characters containing one of a small prespecified set of targets and a number of nontarget (noise) items. He has the task of reporting which target was contained in each array. Estes (I972) has recently presented a model for performance in this and similar visual detection paradigms and applied it with some success to the major findings in the area. According to Estes's model, detection of the target element is mediated by feature-detecting receptive fields. The model assumes that the fields are fairly small, that there are a large number of receptive fields of any given type distributed over the retina, and that the fields are most densely packed at the fovea. The model also assumes mutual inhibitory interactions between receptive fields, and assumes further that the mutual inhibition decreases as distance between the fields increases.Estes (1972) reviewed the three principal varieties of model for visual detection tasks (serial scanning models [e.g., Estes & Taylor, 1964]; parallel limited capacity models [Rumelhart, 1970]; and parallel unlimited capacity models [e.g., Gardner, 1973]), and he found them unable to account for a number of major findings that his feature-detection model can predict. While the three previous models differ in their ability to explain various effects, there is one, which we will call the proximity effect, that none but Estes's model seems able to handle.The proximity effect is found when detectability of the target declines as confusable noise items (nontargets) are moved closer to it in the display. The proximity effect is not a matter of simple lateral masking. First, the effect refers specifically to reduction in visibility of the target caused by noise items that are confusable with the

No abstract

No abstract