A global approach to the design of discrete-time cellular neural networks for associative memories

摘要: In this paper a global design method for associative memories using discrete-time cellular neural networks (DTCNNs) is presented. The proposed synthesis technique enables to realize with several advantageous features. First of all, grey-level as well bipolar images can be stored. Moreover, the approach generates learning and forgetting capabilities. Finally, it possible any kind predetermined interconnection structure. particular, neighbourhoods without line crossings chosen, greatly simplifying VLSI implementation designed DTCNNs. In first part work model multilevel threshold network presented stability analysis carried out basic notions deriving from non-linear dynamical system theory. procedure then developed by means pseudoinversion technique, assuring capabilities DTCNN. use neighbourhood also discussed. Simulation results are reported show capability approach.