The exponential growth of data has exposed the inherent bottlenecks of the von Neumann architecture-specifically its limited computational efficiency and high energy consumption-necessitating an urgent shift toward innovative hardware solutions. Biological perception systems, particularly the human visual system, serve as a premier model for highly integrated, energy-efficient, and multimodal processing, providing a critical blueprint for the future of intelligent computing. Field-effect transistors (FETs) have emerged as a leading platform for visual neuromorphic systems, leveraging their exceptional optoelectronic tunability, mechanical flexibility, and low-power operation. This review provides a comprehensive overview of FET-based visual neuromorphic systems, covering semiconductor material selection, fundamental device architectures, and governing operational principles. Then, the critical role of these devices in emulating biological visual functions is detailed. Finally, the prevailing technical challenges and future development prospects for FET-mediated perception are discussed. This work aims to provide essential insights into the design of the next generation of artificial visual neuromorphic systems and bio-inspired electronics.
Yaqian et al. (Fri,) studied this question.