Handwritten numeral recognition utilizing reservoir computing subject to optoelectronic feedback

Handwritten numeral recognition utilizing optoelectronic feedback reservoir computing is proposed and investigated in this paper. From zero to nine, ten kinds of handwritten numerals have been successfully recognized by numerical simulations. The influence of system parameters on the recognition results has been investigated by adjusting the various parameters of the system. The feedback coefficient has been regulated to change the states of the reservoir to obtain better recognition results. The effects of different input gains and the number of virtual nodes on recognition results have also been studied to obtain the optimal results. It is found that only less training and testing samples are required using the recognition method proposed in this paper.