Analysis and Robust Design for Illusion CNNs

Abstract

The Cellular Neural Network (CNN), proposed by Chua et al in 1988, is a neural network with local connectivity. The CNN has been applied to the fields of image processing, pattern identification, biological visions and so on. The robust design is one of the important issues for the researches in CNN. This paper establishes the local rules of the Muller-Lyer Illusion (MLI) CNN introduced by Chua et. al. The robust design for the MLI CNN is given. A Ponzo Illusion (PI) CNN is introduced. The local rules and the theorem for the PI CNN are set up. Six numerical simulation examples are provided to illustrate the effectiveness of the methodology.