An algebraic construction method for cellular neural networks

An algebraic method for construction of a cellular neural network (CNN) is derived. Using this method, CNNs for connected component detection and the maze problem are given in detail. Because the transition of states is included in the algebraic equations during the construction of the network, the network arrives at a unique solution in either a synchronous or an asynchronous operation process, and the solution is stable. The cellular neural network has important potential applications in such areas as image processing and pattern recognition. Its continuous time feature allows real-time signal processing and its local interconnection feature makes it easy for VLSI implementation.