Optical Instruments, Volume. 46, Issue 6, 64(2024)

OCT retinal images super-resolution reconstruction based on PSRGAN and transfer learning

Minghui CHEN1, Shiyi XU1, Shuting KE1, Yi SHAO2, and Yuquan WU1
Author Affiliations
  • 1Shanghai Engineering Research Center of Interventional Medical Device , University of Shanghai for Science and Technology, Shanghai 200093, China
  • 2Department of Urology, Shanghai General Hospital, Shanghai 200080, China
  • show less
    Figures & Tables(11)
    Structure of the PSRGAN
    The structure of the discriminant in PSRGAN
    The PECA module
    The ECA module
    Dataset image sample
    Comparison of the experimental results of PSRGAN model with and without the PECA attention mechanism
    Image presentation of the experimental results
    Training process presentation of a random image
    Automatic segmentation results of the reconstructed image
    • Table 1. PSRGAN comparison of evaluation indicators on the test set using/without the PECA attention mechanism

      View table
      View in Article

      Table 1. PSRGAN comparison of evaluation indicators on the test set using/without the PECA attention mechanism

      模型RpsnIssimIep
      使用PECA注意力机制30.980.7880.898
      不使用PECA注意力机制30.050.7630.864
    • Table 2. Comparison of PSNR and SSIM for different methods

      View table
      View in Article

      Table 2. Comparison of PSNR and SSIM for different methods

      模型Rpsn/dBIssimIep
      SRCNN27.320.6910.687
      SRGAN29.250.7350.710
      PPECA–SRGAN32.480.835
      PSRGAN32.360.7820.893
      PSRGAN–TL–X-ray32.370.8040.862
      PSRGAN–TL–Flowers33.080.8850.902
    Tools

    Get Citation

    Copy Citation Text

    Minghui CHEN, Shiyi XU, Shuting KE, Yi SHAO, Yuquan WU. OCT retinal images super-resolution reconstruction based on PSRGAN and transfer learning[J]. Optical Instruments, 2024, 46(6): 64

    Download Citation

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

    Category:

    Received: Jan. 24, 2024

    Accepted: --

    Published Online: Jan. 21, 2025

    The Author Email:

    DOI:10.3969/j.issn.1005-5630.202401240011

    Topics