Robust Designs for Gray-Scale Global Connectivity Detection CNN Templates

The cellular neural/nonlinear network (CNN) is a powerful tool for image and video signal processing, as well as robotic and biological visions. Practically, an engineer always hopes to design a CNN that has both universality and robustness. Based on research on the designs for the global connectivity detection (GCD) CNN [Chua, 1997] used in binary pattern, this paper establishes a theorem on robust designs for gray-scale global connectivity detection (GGCD) CNN templates. The theorem provides template parameter inequalities for determining parameter intervals for implementing the GCD functions. As a first example, two gray-scale labyrinth patterns with Gaussian noise are constructed. Using the GGCD, CNN designed by the theorem detects the connectivity of the two labyrinth patterns with gray-scales. In the other three examples, using GGCD CNNs simulate the spreads of an infectious diseases at nonuniform speeds.