Fig. 5: Implementation of pattern classification.
From: Ferroelectric photosensor network: an advanced hardware solution to real-time machine vision
a Two sets of patterns representing letters ‘X’ and ‘T’. b Schematic diagrams showing the operation and the circuit structure of a 1 × 9 FE-PS-NET. The 9 FE-PSs are illuminated following the signals translated from the input pattern. c From left to right: theoretical dimensionless weights, theoretical photoresonsivities scaled from dimensionless weights, actual photoresonsivities right after programming, actual photoresonsivities after pattern classification. The unit of photoresponsivity is mA/W (not shown). d Output currents (I1) during the presentations of different input patterns. The theoretical I1 values are indicated by the “star” symbols. e Neuronal outputs obtained by feeding the output currents to the sigmoid function.
