Fusion of Near-Infrared and UV Light Image via Artificial Visual Neurons Based on Mott Memristor

Multi-band image fusion in biological systems aims to integrate image data from various spectral bands to obtain more comprehensive, accurate, and effective image information. However, developing efficient and low-power artificial vision multi-band image fusion systems inspired by biological vision systems remains a challenge. Here, an artificial visual neuron based on the integration of In₂O₃/PY-IT phototransistor and NbO_x Mott memristor is proposed, which can simultaneously sense optical signals in the UV and near-infrared bands and achieve pulse encoding of different frequencies through light stimulation of different intensities. In addition, the pulse signals encoded by artificial neurons are processed through a Pulse Coupled Neural Network for image fusion, which successfully integrates image information under different lighting scenes and demonstrates the bionic functionality of the artificial vision fusion system. Such artificial visual neurons provide a solid foundation for constructing integrated, functional, and low-power artificial visual systems and serve as building blocks for hardware-based multi-band perception.