Deep-learning-enhanced nonlinear holography in 3D nonlinear photonic crystal

作     者：Zhai, Bohan Chen, Pengcheng Zhang, Zhichao Zhang, Yong 

作者机构：Nanjing Univ Coll Engn & Appl Sci Natl Lab Solid State Microstruct Nanjing 210093 Peoples R China Nanjing Univ Sch Phys Nanjing 210093 Peoples R China Nanjing Univ Collaborat Innovat Ctr Adv Microstruct Nanjing 210093 Peoples R China 

出 版 物：《JOURNAL OF THE OPTICAL SOCIETY OF AMERICA A-OPTICS IMAGE SCIENCE AND VISION》 (J Opt Soc Am A)

年 卷 期：2025年第42卷第3期

页      面：A1-A6页

核心收录：

学科分类：070207[理学-光学] 07[理学] 08[工学] 0803[工学-光学工程] 0702[理学-物理学] 

基　　金：National Natural Science Foundation of China Fundamental Research Funds for the Central Universities [021314380253, 021314380268] Natural Science Foundation of Jiangsu Province [BK20240005, BK20241190] National Postdoctoral Program for Innovative Talents [BX20240157] China Postdoctoral Science Foundation [2024M751366] Jiangsu Funding Program for Excellent Postdoctoral Talent [2023ZB517] 

主　　题：Femtosecond lasers Neural networks Nonlinear photonic crystals Optical encryption Parametric down conversion Sum frequency generation 

摘      要：Nonlinear holography in 3D nonlinear photonic crystals (NPCs) is capable of reconstructing multiple images at newly generated wavelengths, which has attracted increasing interest in optical storage and encryption. However, the storage capacity of nonlinear holography is generally limited by the crosstalk between different channels. In this work, we propose deep-learning-assisted high-capacity nonlinear holography to address this issue. By leveraging the exceptional recognition capability of the ResNet-18 convolutional neural network, the information at multiple channels can be well recognized (i.e., the accuracy reaches 88.1%) in the presence of severe crosstalk (i.e., a bandwidth overlap of 30% between neighboring channels). Importantly, the channel number is doubled in comparison to the previous reports. Our results provide a novel approach to bypass the negative effect of crosstalk for high-security and high-density optical information storage. (c) 2025 Optica Publishing Group. All rights, including for text and data mining (TDM), Artificial Intelligence (AI) training, and similar technologies, are reserved.