Fully‐Printed Transparent InGaO Neuromorphic Visual Systems for Solar‐Blind Bichromatic In‐Sensor Computing

ABSTRACT The accelerating convergence of artificial intelligence urgently necessitates the development of advanced organ‐on‐a‐chip platforms. This study presents a low‐cost transparent multifunctional neuromorphic visual sensor (NVS) utilizing an all‐printed solution method to prepare In 2 O 3 ‐Ga 2 O 3 (IGO) films and silver nanowire (AgNW) electrodes for the first time. The transparent NVS delivers superior optoelectronic properties for neuromorphic computing, inherently supporting in‐sensor image processing, including noise suppression and synaptic weight modulation. The device exhibits unique dual‐band ultraviolet responsiveness at 254 and 365 nm wavelengths, enabling the realization of a bichromatic artificial retina. An in‐sensor computing architecture for handwriting digit recognition was developed. The results showed that IGO‐based NVS exhibits near‐ideal recognition accuracy. Compared with the traditional numerical ANN method, the proposed hardware‐level denoising improved accuracy by over 22% under the highest noise level of 0.9. Besides, the synaptic array maintained 90.2% accuracy even when single‐wavelength synapses failed under noise interference. This breakthrough represents a cost‐effective bio‐inspired approach to visual information processing, demonstrating potential for next‐generation neuromorphic vision systems in artificial intelligence applications.