EQUIVALENT CNN CELL MODELS AND PATTERNS
Abstract
In this paper, canonical isolated CNN cell models are proposed by using implicit differential equations. A number of equivalent but distinct CNN cell models are derived from these canonical models. Almost every known CNN cell model can be classified into one or more groups via constrained conditions. This approach is also applied to discrete-time CNN cell models. Pattern formation mechanisms are investigated from the viewpoint of equivalent templates and genetic algorithms. A strange wave propagation phenomenon in nonuniform CNN cells is also presented in this paper. Finally, chaotic associative memories are proposed.