Extraction of early perceptual structure in dot patterns: integrating region, boundary, and component gestalt

摘要: Abstract This paper presents a computational approach to extracting basic perceptual structure, or the lowest level grouping in dot patterns. The goal is extract segments of dots that group together because their relative locations. are interpreted as belonging interior border segment, being along perceived curve, isolated. To perform grouping, first geometric structure pattern represented terms certain properties Voronoi neighborhoods dots. accomplished through independent modules posses narrow expertise for recognition typical dots, curve and isolated from neighborhoods. results allowed influence change each other so result components satisfy global, Gestalt criteria such smoothness component compactness. Such latera; communication among makes feasible interpretation local manner best meets global expectations. Thus, an integration performed multiple constraints, active at different levels having scopes pattern, infer structure. interpretations well lateral corrections constraint propagation using probabilistic relaxation process. partitioning into tokens. Unlike these size shape addition