A new synthesis procedure for a class of cellular neural networks with space-invariant cloning template

Abstract

This paper presents a new synthesis procedure (design algorithm) for cellular neural networks (CNN's) with a space-invariant cloning template with applications to associative memories. In the present synthesis procedure, the design problem is formulated as a set of linear inequalities, and the inequalities are solved using the well-known perceptron training algorithm. When desired memory patterns are given by a set of bipolar vectors, it is guaranteed that a cellular neural network with a space-invariant cloning template can be designed using the design algorithm developed herein. An algorithm is also provided to design CNN's with space-invariant cloning templates and with symmetric connection matrices to guarantee the global stability of the network. Two specific examples are included to demonstrate the applicability of the methodology developed herein.Index Terms-Associative memories, cellular neural networks, perceptron training.

Detection of microcalcification in digitized mammograms with multistable cellular neural networks using a new image enhancement method: automated lesion intensity enhancer (ALIE)

“…The training process of CNNs aims to obtain appropriate template matrices for the desired output. There are many varieties of CNN templates, like adaptive templates, linear templates, and cloning templates [14][15][16][17][18][19][20][21].…”