IMPLEMENTATION OF INTER-PATTERN ASSOCIATION NEURAL NETWORKS FOR AUTOMATIC TARGET RECOGNITION USING NOVEL ELECTRON TRAPPING MATERIALS

WE PROPOSE USING ELECTRON TRAPPING (ETTM) MATERIALS FOR THE IMPLEMENTATION OF THE INTERCONNECTION WEIGHT MATRIX (IWM) IN AN OF USING ET MATERIALS IS THAT THEY OFFER A LARGE STORAGE CAPACITY, BY WHICH 1.6X10 TO THE 10 POWER RECONFIGURABLE INTERCONNECTIONS ARE POSSIBLE. THE PROPOSED ET ELECTRO-OPTIC NEURAL NETWORK CAN PERFORM 2X10 TO THE 3RD POWER ASSOCIATIONS PER SECOND, AND RECOGNIZE 1.5X10 TO THE 3RD POWER TARGETS, WITH SHIFT, SCALE AND 3D ROTATION INVARIANCES.