A synergetics approach to image processing in cellular neural networks

The nonlinear array dynamics of the simplest Cellular Neural Network (CNN) will be shown to exhibit the basic phenomena of “synergetics”. Synergetics has been used to explain the emergence of large scale organization in many systems of physics, chemistry, and biology. Here, it is suggested that the computation performed by these systems could be emulated in a natural way on the CNN to perform useful image processing tasks. Namely, the physical phenomena can be interpreted in terms of linear spatial filtering and morphological constraints. The classical CNN “averaging” template will be interpreted in this light and, informed by the synergetics interpretation, improved for a particular example