Acta Optica Sinica, Volume. 43, Issue 20, 2010002(2023)

Unsupervised Denoising of Retinal OCT Images Based on Deep Learning

Guangyi Wu, Zhuoqun Yuan, and Yanmei Liang*
Author Affiliations
  • Institute of Modern Optics, Nankai University, Tianjin Key Laboratory of Micro-Scale Optical Information Science and Technology, Tianjin 300350, China
  • show less
    Figures & Tables(9)
    Network structure of DRSA-Net
    Flow chart of retinal OCT image unsupervised training experiment
    Noise reduction in retinal OCT images. (a) Original noisy image; (b) denoised images of BM3D; (c) denoised image of DnCNN-N2N; (d) 5-frame average ground truth images; (e) denoised image of U-Net-N2N; (f) denoised image of Ours-N2N
    Noise reduction results of supervised learning and unsupervised learning retinal OCT images. (a) Original noisy image; (b) denoised image of DnCNN-N2C; (c) denoised image of U-Net-N2C; (d) denoised image of Ours-N2C; (e) 5-frame average ground truth image; (f) denoised image of DnCNN-N2N; (g) denoised image of U-Net-N2N; (h) denoised image of Ours-N2N
    Unsupervised learning generalization ability test. (a) Original noisy image; (b) denoised images of BM3D; (c) denoised image of DnCNN-N2N; (d) 40-frame average ground truth images; (e) denoised image of U-Net-N2N; (f) denoised image of Ours-N2N
    Comparison of generalization ability between supervised and unsupervised learning. (a) Denoised image of DnCNN-N2C; (b) denoised image of U-Net-N2C; (c) denoised image of Ours-N2C; (d) denoised image of DnCNN-N2N; (e) denoised image of U-Net-N2N; (f) denoised image of Ours-N2N
    • Table 1. Results of supervised learning and unsupervised learning denoising numerical evaluation

      View table

      Table 1. Results of supervised learning and unsupervised learning denoising numerical evaluation

      MethodPSNRSSIMEPIENLTime /s
      BaselineNoisy19.073±0.0650.391±0.00214.938±1.693
      BM3D23.937±0.1810.253±0.1010.281±0.051434.819±111.647127
      Supervised learning modelDnCNN25.418±0.2130.506±0.0070.295±0.023220.663±14.4720.48
      U-Net24.852±0.3220.498±0.0060.304±0.024255.427±20.6420.70
      Ours25.447±0.1910.494±0.0070.312±0.025279.760±27.9460.53
      Unsupervised N2N modelDnCNN24.234±0.1830.284±0.0070.242±0.027768.128±137.6060.48
      U-Net24.543±0.2120.287±0.0080.243±0.0281601.956±573.3280.70
      Ours24.582±0.2250.289±0.0080.262±0.0261304.384±466.9830.53
    • Table 2. Results of supervised learning and unsupervised learning denoising numerical evaluation

      View table

      Table 2. Results of supervised learning and unsupervised learning denoising numerical evaluation

      MethodPSNRSSIMEPIENLTime /s
      BaselineNoisy18.193±0.2730.134±0.02414.074±3.681
      BM3D28.035±1.4570.550±0.0440.268±0.035418.037±106.729256
      Supervised learning modelDnCNN28.699±1.3010.554±0.0360.277±0.027216.478±37.8980.79
      U-Net27.956±1.1860.569±0.0350.287±0.028241.361±31.4991.04
      Ours29.317±1.4190.593±0.0340.293±0.030262.771±34.7620.92
      Unsupervised N2N modelDnCNN29.673±1.1950.665±0.0120.166±0.037628.151±182.5290.79
      U-Net30.950±1.5650.702±0.0120.176±0.0361266.897±338.5431.04
      Ours31.172±1.7060.706±0.0130.194±0.0351029.639±220.7140.92
    • Table 3. Ablation experimental results of different modules of network

      View table

      Table 3. Ablation experimental results of different modules of network

      ModulePSNRSSIMEPIENLTime/s
      DEB+GAB+RB28.962±1.6320.703±0.0150.131±0.031159.334±13.5020.73
      LSAB+GAB+RB28.737±1.2640.613±0.0260.133±0.072879.870±207.9820.69
      LSAB+DEB+RB30.301±1.4460.647±0.0120.150±0.035903.572±343.2170.73
      LSAB+DEB+GAB30.802±1.4270.701±0.0120.127±0.033937.421±225.0730.75
      LSAB+DEB+GAB+RB31.172±1.7060.706±0.0130.194±0.0351029.639±220.7140.92
    Tools

    Get Citation

    Copy Citation Text

    Guangyi Wu, Zhuoqun Yuan, Yanmei Liang. Unsupervised Denoising of Retinal OCT Images Based on Deep Learning[J]. Acta Optica Sinica, 2023, 43(20): 2010002

    Download Citation

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

    Category: Image Processing

    Received: Mar. 27, 2023

    Accepted: May. 6, 2023

    Published Online: Oct. 23, 2023

    The Author Email: Yanmei Liang (ymliang@nankai.edu.cn)

    DOI:10.3788/AOS230720

    Topics