Export citation

Export citation

Choose format for download:

Download Citation

    Triple-key optical encryption system using ultra-sparse sampling of orbital angular momentum speckles

    Junlei Zhou1, Yaling Yin1, Qi Chu1, Chaoxiu Guo1, Quanli Gu2, and Yong Xia1,3,4,*

    • *Contact author: yxia@phy.ecnu.edu.cn

    Phys. Rev. Applied 26, 024074 – Published 26 August, 2026

    DOI: https://doi.org/10.1103/twcq-vxmx

    Abstract

    We demonstrate a triple-key optical encryption system that integrates orbital angular momentum (OAM) multiplexing with physical scattering and ultra-sparse sampling. At its core is an adaptive ResNeXt-SPD decoder that achieves 99.28% accuracy in decoding 24-bit OAM superposition states with only 64 pixels—a sampling point density of 0.13%. The system implements a cross-domain security architecture comprising three keys: a physical key (an unclonable scattering medium), a digital key (a sparse sampling pattern), and a computational key (trained network weights). We validate the sparse sampling performance in the full architecture, explore the sampling threshold under constrained networks, and investigate the criticality of each security key. Security analysis confirms that the absence or mismatch of any single key renders the system inoperative. Substituting the physical key drops accuracy to approximately 50%. One pixel shift in the digital key reduces accuracy from 99.28% to approximately 50%. Untrained model weights yield performance equivalent to random guessing. End-to-end image transmission under extreme sparsity maintains high reconstruction quality. This work advances OAM-based secure communication by achieving the lowest sampling-density and multilayer cryptographic protection simultaneously.

    Physics Subject Headings (PhySH)

    Authorization Required

    We need you to provide your credentials before accessing this content.

    References (Subscription Required)

    Outline

    Information

    Sign In to Your Journals Account

    Filter

    Filter

    Article Lookup

    Enter a citation