A new compact neuron-bipolar cellular neural network structure with adjustable neighborhood layers and high integration level

Abstract

A new compact structure called the neuron-bipolar cellular neural network is proposed and analyzed. In the new structure, the parasitic PNP bipolar junction transistor in the CMOS process is used to implement the neuron whereas the coupling MOS resistor is used to realize the synapse weights among various neurons. It has the advantages of compact structure and small chip size. The neuron-bipolar CNN with single neighborhood layer r=1 has been successfully applied to the noise removing operation on images. Moreover, the proposed neuron-bipolar CNN can easily realize the multi-layer neighborhood structure. Thus the proposed neuron-bipolar CNN has a great potential in the VLSI implementation of neural networks.