Design for Robustness Black-White Color Interconversion CNN

The cellular neural/nonlinear network (CNN) is a powerful tool for image and video signal processing, robotic and biological visions. The robust designs for CNN templates are important issue for the practical applications of the CNN. This paper introduces a new kind of CNNs called black-white color interconversion CNN which can turn pixels from white to black or from black to white. As a main contribution, we first establish one theorem which provides parameter inequalities for determining parameter intervals for implementing prescribed image processing functions. Then we found that different template parameter require different processing time and the parameters with more processing time can simulate some nature phenomena, which are very interesting and worthy of future study. Simulation result shows the effectiveness of the proposed methodology.