AREA-EFFICIENT IMPLEMENTATIONS OF FIXED-TEMPLATE CNNS

Implementations of fixed-template Cellular Neural Networks (CNN's) wit h reduced circuit complexity are presented, Considerable improvements in area without performance degradation have been obtained by: 1) usin g single-polarity signals that reduce the number of transistors requir ed for signal replication and to generate the pseudo-linear output fun ction; 2) using simple current-mode circuits to implement the output p seudo-linear function; and 3) searching for network parameter configur ations that solve a particular application using the proposed circuit implementation with less hardware complexity. Experimental results for a CCD-CNN chip prototype with a density of 230 cells per millimeter-s quared (mm(2)) are also reported.