Large neighbourhood template implementation in continuous-time cellular neural networks with physical connectivity of r=1

A new approach to the implementation of large neighbourhood (r>1) templates in continuous-time cellular neural networks (CT-CNNs) is presented. The new method preserves r=1 physical connectivity (i.e. local interconnections only) and is thus attractive for VLSI realisations of CT-CNNs with r>1. The method is based on the assumption that the circuit transients are monotonic, that the pixel values at the inputs are binary valued (/spl plusmn/1) and that the values of the state voltage variables are constrained to /spl plusmn/1. Simulation results are presented to confirm the viability of the proposed method.