IGZO/InHfO $_{\textit{x}}$ Nanowires/IGZO Phototransistor With Persistent Photoconductivity Effect for Intelligent Visual Perception Application

Traditional computing systems necessitate frequent data transfers and communication between physically separated chips for acquisition, storage, and computation. This challenge is further intensified during the complex training of recurrent neural networks (RNNs) for edge device applications. Neuromorphic computing systems based on artificial synapses (AS) demonstrate considerable potential in the field of brain-inspired information processing. In this work, we report an intelligent visual perception system utilizing IGZO/InHfO
$_{\textit{x}}$
nanowires (NWs)/IGZO-based neuromorphic phototransistors. We further exploited the device’s ability to process temporal visual information by leveraging its synaptic plasticity. By incorporating dynamic reservoir computing (RC) and a dedicated readout function, we achieved high-accuracy recognition of various standard digit categories. This work demonstrates a multiclassification hardware circuit built upon the intrinsic nonlinear sustained light response of the neuromorphic photodetector. The ZYNQ7020 development board, integrating the software programmability of a processor with the hardware programmability of a field-programmable gate array (FPGA), provides a more efficient and flexible hardware platform for neuromorphic computation based on these synaptic devices.