Chinese Journal of Lasers, Volume. 47, Issue 12, 1206005(2020)

Image Reconstruction of Multimode Fiber Scattering Media Based on Deep Learning

Meng Lu1、*, Hu Haifeng1,2, Hu Jinzhou1, Bu Sihang1, and Gao Han1
Author Affiliations
  • 1College of Information Science and Engineering, Northeastern University, Shenyang, Liaoning 110004, China
  • 2School of Optical-Electrical and Computer Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China
  • show less
    Figures & Tables(13)
    Design of optical experiment platform
    Overall network structure. (a) Conv block; (b) dense block; (c) DenseUnet; (d) DCNN
    ReLU nonlinear activation function
    Physical diagram of multimode optical fiber imaging system
    Digital images in imaging system performance test experiment
    Results of reconstruction and classification
    Classification accuracy of training set and validation set for reconstructed images under 0.1 m fiber
    Reconstruction accuracy under different fiber lengths. (a) 0.1 m; (b) 10 m
    Classification accuracy of different fiber lengths in training set and validation set. (a) 0.1 m; (b) 10 m
    Results of speckle reconstruction under different fiber lengths
    Comparison of reconstruction results of different network models. (a) Speckle images output from multimode fiber; (b) reconstruction results of DenseUnet; (c) reconstruction results of traditional Unet; (d) original image (ground truth)
    • Table 1. Classification accuracy of different fiber lengths using two input modes

      View table

      Table 1. Classification accuracy of different fiber lengths using two input modes

      Fiberlength /mModelClassification accuracy /%
      SpeckleimageReconstructedimage
      0.1DCNN91.5±0.498.2±0.2
      1DCNN90.2±0.397.6±0.2
      10DCNN86.6±0.697.2±0.5
      20DCNN76.4±1.197.0±0.4
    • Table 2. Reconstruction quality of different models

      View table

      Table 2. Reconstruction quality of different models

      Fiber length /mIndexDenseUnetUnetDenseUnet-
      SSIM0.8980.7190.658
      0.1PSNR41.94334.02211.978
      Classification accuracy /%98.2095.3385.25
      SSIM0.7900.7120.622
      1PSNR36.44028.10811.402
      Classification accuracy /%97.6092.6779.67
      SSIM0.7800.5400.594
      10PSNR36.07212.12511.344
      Classification accuracy /%97.2073.3368.72
      SSIM0.7600.4220.517
      20PSNR35.94311.91713.002
      Classification accuracy /%975368.67
    Tools

    Get Citation

    Copy Citation Text

    Meng Lu, Hu Haifeng, Hu Jinzhou, Bu Sihang, Gao Han. Image Reconstruction of Multimode Fiber Scattering Media Based on Deep Learning[J]. Chinese Journal of Lasers, 2020, 47(12): 1206005

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Fiber optics and optical communication

    Received: Jun. 22, 2020

    Accepted: --

    Published Online: Nov. 17, 2020

    The Author Email: Lu Meng (menglu@ise.neu.edu.cn)

    DOI:10.3788/CJL202047.1206005

    Topics