Deep-learning-based recognition of multi-singularity structured light

Abstract Structured light with customized topological patterns inspires diverse classical and quantum investigations underpinned by accurate detection techniques. However, the current schemes are limited to vortex beams a simple phase singularity. The precise recognition of general structured multiple singularities remains elusive. Here, we report deep learning (DL) framework that can unveil multi-singularity structures in an end-to-end manner, after feeding only two intensity upon beam propagation. By outputting directly, rich intuitive information twisted photons is unleashed. DL toolbox also acquire phases Laguerre–Gaussian (LG) modes single singularity other objects likewise. Enabled this platform, phase-based optical secret sharing (OSS) protocol proposed, which based on more class than conventional LG beams. OSS features strong security, wealthy state space, convenient intensity-based measurements. This study opens new avenues for large-capacity communications, laser mode analysis, microscopy, Bose–Einstein condensates characterization, etc.