Laser & Optoelectronics Progress, Volume. 61, Issue 8, 0809001(2024)

Color Hologram Reconstruction Based on Deep Learning

Juntong Liu1, Jinbin Gui1,2、*, Aishuai Chen1, Xiandong Ma1, and Xianfei Hu1
Author Affiliations
  • 1Science of Faculty, Kunming University of Science and Technology, Kunming 650500, Yunnan , China
  • 2Yunnan Provincial Key Laboratory of Modern Information Optics, Kunming University of Science and Technology, Kunming 650500, Yunnan , China
  • show less
    Figures & Tables(12)
    Hologram recording optical path map
    Zero filling reconstruction of color hologram. (a) Green light reconstruction image; (b) blue light reconstruction image; (c) red light reconstruction image (d) synthetic color holographic reconstruction image
    Schematic diagram of CHR-Net structure
    Deep supervision mechanism
    Example of dataset production
    Network error curve
    Output results from three branches under deep supervision; (a) Original image; (b) out2; (c) out1; (d) out
    Comparison of reconstruction effects of different algorithms. (a) Original images; (b) S-FFT reconstruction images; (c) U-Net reconstruction images; (d) CHR-Net reconstruction images
    Verification of Fresnel hologram reconstruction from actual shooting. (a) Original image; (b) S-FFT reconstruction image; (c) U-Net reconstruction image; (d) CHR-Net reconstruction image
    • Table 1. Reconstructed image PSNR value

      View table

      Table 1. Reconstructed image PSNR value

      ImageS-FFTU-NetCHR-Net
      Image124.025127.216635.4487
      Image222.199918.810224.6732
      Image322.589217.743033.0315
      Image419.215822.751826.8650
      Average22.007521.630430.0046
    • Table 2. Reconstructed image SSIM value

      View table

      Table 2. Reconstructed image SSIM value

      ImageS-FFTU-NetCHR-Net
      Image10.60900.94760.9878
      Image20.34350.68710.6536
      Image30.32650.80660.9546
      Image40.45270.94450.9692
      Average0.43290.84650.8913
    • Table 3. Overall network performance evaluation

      View table

      Table 3. Overall network performance evaluation

      IndexS-FFTU-NetCHR-Net
      PSNR22.345222.286628.6838
      SSIM0.46720.88530.9104
    Tools

    Get Citation

    Copy Citation Text

    Juntong Liu, Jinbin Gui, Aishuai Chen, Xiandong Ma, Xianfei Hu. Color Hologram Reconstruction Based on Deep Learning[J]. Laser & Optoelectronics Progress, 2024, 61(8): 0809001

    Download Citation

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

    Category: Holography

    Received: Jun. 8, 2023

    Accepted: Jul. 31, 2023

    Published Online: Mar. 5, 2024

    The Author Email: Gui Jinbin (jinbingui@163.com)

    DOI:10.3788/LOP231492

    Topics