Perturbation model for letter identification.

Abstract

A model is presented which provides a reasonably good quantitative account of the data from a scries of tachistoscopic report experiments in which retinal locus, spacing, and array position were systematically varied. Potential extensions of the model to other paradigms are also suggested. The main assumptions of the model are that features are extracted from visual displays in unlimited parallel channels as a function of retinal locus; that a processor, working serially, attempts to identify letters from the extracted features; and that during processing, perturbations occur in the spatial ordering of the features.

Illusory conjunctions and the cerebral hemispheres

“…However, we did not observe distance (or size) effects on ICs, as measured by d' scores. This result is inconsistent with both Prinzmetal and Keysar's (1989) and Wolford's (1975) proposal that ICs should occur more readily between items that are close together than between items that are further apart. Yet, it corroborates Treisman and Schmidt's (1982) observations that spatial separation has little effect on the frequency of ICs (see more precise discussions in Cohen & Ivry, 1989, 1991.…”

The role of structural redundancy in the perception of visual targets

“…Some very interesting work along this line has been reported by Banks and Prinzmetal (1972). They have shown that "higher order" forms produced by arrangements of smaller letters influence the interactions between these letters beyond that attributable to simple proximity (e.g., Wolford, 1975). Again, whether the "organization" of higher order forms precedes, follows, or is simultaneous with the processing of lower order forms seems entirely open at this point.…”

“…There is considerable evidence that letters located close together are not processed independently (e.g., Wolford, 1975). The two large letters in the present experiment were purposely separated by over 4°in order to minimize such interactions.…”

Principles of feature integration in visual perception

“…Both these failures of feature abstraction will be called/eoture errors (Treisman & Gelade, 1980). Recently several theorists have postulated another type of perceptual error that is not related to feature abstraction, but rather to a failure to appropriately join the features that have been correctly abstracted (Treisman, Sykes, & Gelade, 1977;Wolford, 1975). Two examples of this type of error are as follows.…”