MOS implementation of oscillatory neural Ellias-Grossberg model

Abstract

The oscillatory neural networks are kind of networks that try to model the behavior of the biological neurons in the part of the nervous system responsible of the vision processing. These networks are developing fast in the applications of image processing. Although they consume a lot of processing time by computer simulations, their implementation on integrated circuit is still in the early stages. We develop a MOS transistor circuit that implements Ellias-Grossberg (EG) oscillatory model. This development came on three stages: the behavioral simulation stage, the pre-circuit level stage and the transistor level implementation stage. The behavior of the three levels of implementation is compared.