Continuous-time and discrete-time cellular neural networks

Abstract

Publisher Summary This chapter discusses the design of cellular neural network (CNN) for different local and global tasks. Many features of CNN are based on state-of the-art VLSI technology. CNN promises tremendous parallel computation power and local connectedness, which is, in nature, a suitable tool for image processing. The chapter presents a detailed analysis of different types of CNNs for the same task class and discusses several aspects of applications of CNNs to image processing. The organization of references is based on their contributions to the applications of CNN to image processing problems. The chapter presents a set of standard boundary conditions. Some CNNs may fail to work correctly without specified boundary conditions, while others are independent of boundary conditions. Propagating CNNs are usually sensitive to boundary conditions. Boundary conditions were neglected by many authors, resulting occasionally in misleading results.