Assignment of Figural Side to Contours Based on Symmetry, Parallelism, and Convexity

摘要

We propose a neural network model for the figure-ground organization based on spatial arrangement of contours such as parallelism, symmetry, and on contour convexity. All of them have been manifested as effective factors for figure-ground organization by psychological studies. In order to measure parallelism and symmetry, spatially separated distant contours have to be corresponded. Our model process them by local detectors embedded in hierarchical architecture of network in which image data is pyramidally encoded. We tested our model by computer simulation and succeeded to mimic some human perceptions.