Deep learning image transmission through a multimode fiber based on a small training dataset.

Journal: Optics express
Published Date:

Abstract

An improved deep neural network incorporating attention mechanism and DSSIM loss function (AM_U_Net) is used to recover input images with speckles transmitted through a multimode fiber (MMF). The network is trained on a relatively small dataset and demonstrates an optimal reconstruction ability and generalization ability. Furthermore, a bimodal fusion method is developed based on S polarization and P polarization speckles, greatly improving the recognition accuracy. These findings prove that AM_U_Net has remarkable capabilities for information recovery and transfer learning and good tolerance and robustness under different MMF transmission conditions, indicating its significant application potential in medical imaging and secure communication.

Authors

  • Binbin Song
  • Chang Jin
  • Jixuan Wu
  • Wei Lin
    Department of Geriatric Rehabilitation, Jiangbin Hospital, Nanning, China.
  • Bo Liu
    Wuhan United Imaging Healthcare Surgical Technology Co., Ltd., Wuhan, China.
  • Wei Huang
    Shaanxi Institute of Flexible Electronics, Northwestern Polytechnical University, 710072 Xi'an, China.
  • Shengyong Chen