Real time CCD-based neural network system for pattern recognition applications

Abstract

A generic NNC (neural network classifier) capable of providing 1.9 billion programmable connections per second is described. Applications for these generic processors include image and speech recognition as well as sonar signal identification. To demonstrate the modularity and flexibility of the CCD (charge coupled device) NNCs, two generic multilayer system-level boards capable of both feedforward and feedback nets are presented. The boards demonstrate multiple LL NN chips in an adaptable, reconfigurable, expandable multipurpose system design. Although only two examples are demonstrated, the extension to larger and more complicated networks using multiple NN devices as building blocks is straightforward.<>