Journals Articles Conferences News About CLP
Submit a paper Email Alert
Search
Search
Articles
Journals
News
Advanced Search
Top Searches
2D MaterialsTransformation opticsQuantum PhotonicsSmall lasersSemiconductor UV PhotonicsSupersymmetric microring laser arrays

Chinese Journal of Liquid Crystals and Displays, Volume. 38, Issue 9, 1224(2023)

Blind super-resolution model based on degradation-aware

Jian-feng CAI and Nian-de JIANG*
Author Affiliations
  • School of Information Engineering，East China University of Technology，Nanchang 330013，China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (15)
    Equations (0)
    References (27)
    Tools
    Paper Information
    Topics
    Figures & Tables(15)
    Overall structure diagram based on degradation aware

    Fig. 1. Overall structure diagram based on degradation aware

    Download full size
    View in Article
    Degenerate-aware block

    Fig. 2. Degenerate-aware block

    Download full size
    View in Article
    Degenerate representation cross attention block

    Fig. 3. Degenerate representation cross attention block

    Download full size
    View in Article
    Layer feature reuse block

    Fig. 4. Layer feature reuse block

    Download full size
    View in Article
    Attention map of different layers

    Fig. 5. Attention map of different layers

    Download full size
    View in Article
    Comparison of “bird” recovery results in Set5

    Fig. 6. Comparison of “bird” recovery results in Set5

    Download full size
    View in Article
    Comparison of “pepper” recovery results in Set14

    Fig. 7. Comparison of “pepper” recovery results in Set14

    Download full size
    View in Article
    Comparison of “ppt3” recovery results in Set14

    Fig. 8. Comparison of “ppt3” recovery results in Set14

    Download full size
    View in Article
    Comparison chart 1 of the real image magnified 4 times

    Fig. 9. Comparison chart 1 of the real image magnified 4 times

    Download full size
    View in Article
    Comparison chart 2 of the real image magnified 4 times

    Fig. 10. Comparison chart 2 of the real image magnified 4 times

    Download full size
    View in Article
    Comparison chart 3 of the real image magnified 4 times

    Fig. 11. Comparison chart 3 of the real image magnified 4 times

    Download full size
    View in Article

    • Table 1. Experimental results of different module combinations

      View table
      View in Article

      Table 1. Experimental results of different module combinations

      模型Set5Set14
      PSNR/dBSSIMPSNR/dBSSIM
      Baseline31.140.880 527.580.752 5
      Baseline+ca31.130.881 627.780.755 8
      Ours31.170.881 227.830.757 2

    • Table 2. Test results of different SR algorithms at 2X magnification factor

      View table
      View in Article

      Table 2. Test results of different SR algorithms at 2X magnification factor

      Method ×2Set5Set14Urban100BSD100
      PSNR/SSIMPSNR/SSIMPSNR/SSIMPSNR/SSIM
      Bicubic28.32/0.811 626.42/0.723 523.12/0.664 825.93/0.672 7
      RCAN28.37/0.824 926.22/0.735 023.46/0.677 526.06/0.683 3
      HAN28.71/0.820 826.94/0.738 623.58/0.682 226.21/0.685 4
      DASR30.35/0.874 428.23/0.787 425.14/0.750 027.48/0.745 5
      Ours33.37/0.927 029.75/0.838 826.52/0.811 028.80/0.794 5

    • Table 3. Test results of different SR algorithms at 4X magnification factor

      View table
      View in Article

      Table 3. Test results of different SR algorithms at 4X magnification factor

      Method ×4Set5Set14Urban100BSD100
      PSNR/SSIMPSNR/SSIMPSNR/SSIMPSNR/SSIM
      Bicubic26.66/0.754 924.97/0.651 522.08/0.598 524.97/0.614 0
      RCAN25.99/0.751 324.88/0.650 223.26/0.654 525.53/0.645 0
      HAN27.64/0.792 725.97/0.689 423.04/0.655 425.59/0.647 0
      IKC29.87/0.822 627.03/0.696 824.50/0.699 826.75/0.672 9
      DASR30.85/0.876 527.72/0.751 524.71/0.732 027.03/0.708 0
      Ours31.17/0.881 227.83/0.757 224.73/0.735 027.10/0.713 1

    • Table 4. Quantitative comparison of the models on RealSR

      View table
      View in Article

      Table 4. Quantitative comparison of the models on RealSR

      MethodRealSR PSNR/SSIM
      RCAN25.97/0.779 4
      HAN25.97/0.779 2
      DASR26.09/0.793 1
      IKC25.81/0.750 7
      Ours26.13/0.793 3
    Tools

    Get Citation

    Copy Citation Text

    Jian-feng CAI, Nian-de JIANG. Blind super-resolution model based on degradation-aware[J]. Chinese Journal of Liquid Crystals and Displays, 2023, 38(9): 1224

    Download Citation

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

    Category: Research Articles

    Received: Nov. 18, 2022

    Accepted: --

    Published Online: Sep. 19, 2023

    The Author Email: Nian-de JIANG (cjnd@163.com)

    DOI:10.37188/CJLCD.2022-0385

    Topics