ABSTRACT
This chapter describes a new type of system that is capable of automatic figure-ground separation. This process separates scenic figures whose emergent boundary segmentations surround a connected region. The chapter contributes to the development of a self-organizing neural network architecture for invariant pattern recognition in a cluttered environment. It reviews a method for figure-ground separation that uses combinations of laser radars, or related artificial detectors. The chapter also reviews a method that arises in a neural network model of biological vision that has been called FACADE theory. It discusses the FBF model that is capable of separating connected figures from their backgrounds in response either to monochromatic images, such as a gray-scale photograph, or from images derived from multiple detectors. The chapter provides an intuitive description of network stages and their effects, and describes network equations and parameters.
