Demonstrates enhanced motion perception in bio-inspired sensors, highlighting potential for real-time visual adaptation.
Conventional machine visions, which rely on complex algorithms and hardware that separate image acquisition from processing, face challenges with latency and power consumption in real‐time information perception. While neuromorphic bio‐inspired devices present a promising alternative, their typical dependence on charge trapping and de‐trapping mechanisms often constrains response speed and limits their applicability in low‐power, real‐time motion perception scenarios. In this study, we develop a functionally retina‐inspired optoelectronic sensor based on two‐dimensional γ‐InSe, where engineered interlayer sliding enhances spontaneous polarization. Leveraging the photo‐pyroelectric effect under zero‐bias operation, the sensor emulates the visual adaptation functions of the snake retina. Configured as a 5 × 5 array, it achieves integrated spatiotemporal perception over a broad spectral range (400–1100 nm), enabling all‐day motion tracking. The sensor exhibits a rapid response time of 37.4 ms, a wide dynamic range of 103 dB, and reliable decoding of motion trajectories under varying illumination—all in a self‐powered mode. When integrated with a neural network, the system achieves simultaneous inference of multiple target parameters, including direction, velocity, ambient lighting conditions, and rotation, achieving trajectory recognition accuracy exceeding 99.6%.
No takes yet. Share an insight, caveat, or question.
Wang et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: