Fast Binocular Visible‐Infrared Computational Imaging for Palmprint and Vein Detection Using Narrow Bandgap Perovskite Photodetectors

Abstract Noninvasive/contact palm‐vein recognition utilizes near‐infrared (NIR) illumination to highlight subcutaneous venous patterns, enabling deep‐tissue imaging with high resolution and improved contrast, binocular dual‐spectral imaging representing a breakthrough in biometric identification. Narrow bandgap Sn‐Pb perovskites have recently emerged as potential sensing materials that may outperform conventional NIR counterparts. However, the instability, nonuniform large‐scale pixel array fabrication, as well as crosstalk greatly hinders high‐quality imaging. Here, a novel binocular imaging system that combines fast self‐powered visible (VIS)/NIR perovskite photodetectors with advanced computational algorithms is proposed. To date, it features exceptional imaging speed of 2.9 µs, a record‐low detection limit of 7.64 nW cm − 2 . Meanwhile, the NIR photodetectors (PDs) are applicable to convert transmission NIR light to photovoltage. By integrating the anti‐scattering merit of single‐pixel imaging with deep penetration of infrared light in tissue, high‐performance binocular computational imaging in diffuse reflection and projection modes has been achieved for the first time. Furthermore, fast noncontact palmprint and palm‐vein recognition imaging are successfully demonstrated within 1 s. This work offers a promising new direction for perovskite‐based NIR imaging, advancing the development of more reliable and detailed imaging capabilities.