Edge Detection Using Young's Reaction-Diffusion Cellular Automaton with Nonlinear Function

Abstract

The two-dimensional cellular automaton (CA) model proposed by D. A. Young is known to generate spotted and labyrinthine patterns. Our previous study extended this model and found that it can perform image processing tasks such as edge detection and skeletonization depending on the parameters. We also identified parameter conditions that enable edge detection, although this was limited to cases where a linear function controls the CA local rules. In this study, we hypothesized and verified that employing a nonlinear control function makes these conditions easier to satisfy. For comparison, we modified only a portion of the linear control function to make it nonlinear. As a result, we found that edge detection can be achieved with a nonlinear function even in cases where the original linear function fails.