A Flexible, Broadband Optoelectronic Memristor for Neuromorphic Visual Perception and Intelligent Image Processing

Abstract The compartmentalization of sensing, storage, and computing functions in conventional artificial vision systems typically results in high‐energy consumption and data transfer bottlenecks, which impede the development of next‐generation artificial vision systems. Here, a highly flexible bio‐inspired neuromorphic optoelectronic‐memristor array based on aluminum‐doped zinc oxide (AZO) is achieved, demonstrating broadband optical response (UV‐RGB), reconfigurable optoelectronic synapses, and super‐resolution image reconstruction. By engineering a vertically integrated O D ‐AZO/O R ‐AZO homojunction structure, the device possesses dynamic conductance modulation through synergistic optical and electrical inputs, supporting comprehensive synaptic behaviors including short‐term memory, long‐term memory, and paired‐pulse facilitation while also exhibiting excellent flexibility, high endurance, and stable retention. When configured as neuromorphic visual arrays, the system exhibits glorious image perception and optical information storage. Furthermore, a super‐resolution reconstruction neural network is constructed based on precisely tuned memristor conductance. The system exhibits an average peak signal‐to‐noise ratio of 25.01 dB and an average structural similarity index of 0.709. Systematic characterization reveals that oxygen vacancy gradient engineering in the AZO matrix facilitates stable tunneling current while maintaining mechanical integrity under bending conditions. This work provides a materials design paradigm for developing multifunctional neuromorphic vision platforms compatible with flexible electronics and edge computing applications.