PERCEIVING TOPOLOGICAL-STRUCTURE OF 2-D PATTERNS

Four experiments investigated observers' sensitivity to the topologica l structure of visual stimuli. Three factors were taken to capture the topological structure of 2-D patterns: The number of disconnected com ponents, the number of holes (connections), and inclusion relationship s. If studied in isolation, any given topological property is typicall y confounded with the presence of particular features such as line ter minations and contour length, or with Gestalt principles of perceptual organization. We went beyond existing studies and attempted to system atically remove potential confounds from the stimulus displays. Result s showed that processing speeds for two-dimensional patterns are a fun ction of their topological properties, The more patterns differ in the ir topological structure the easier they can be discriminated. Not onl y do all three topological factors contribute to pattern discriminabil ity, they also can be combined to provide an overall measure of struct ural complexity. Forced choice comparison techniques agreed well with similarity judgments. Topological structure thus contributed to discri minability above and beyond many confounding variables. Claims suggest ing a general topological analyzer in visual processing are discussed (C) 1998 Elsevier Science B.V. All rights reserved.