Optics and Precision Engineering, Volume. 31, Issue 10, 1509(2023)

Full reference image quality assessment based on color appearance-based phase consistency

Benchi JIANG... Shilei BIAN, Chenyang SHI* and Lulu WU |Show fewer author(s)
Author Affiliations
  • School of Artificial Intelligence, Anhui Polytechnic University, Wuhu241000, China
  • show less
    Figures & Tables(11)
    Performance comparison chart of CPC algorithm and traditional PC algorithm
    Comparison of image contrast and chromaticity
    Characteristic similarity of images
    Analytic flow of the proposed model
    SROCC of different C1 and C values on four benchmark databases
    SROCC of different w1 and w2values on four benchmark databases
    Scatter diagram of subjective MOS and model prediction scores on TID 2013 database
    • Table 1. IQA benchmark database

      View table
      View in Article

      Table 1. IQA benchmark database

      DatabaseSource imagesDistorted imagesDistortion typesObservers
      TID2013253 00024971
      TID2008251 70017838
      CSIQ30866635
      LIVE297795161
    • Table 2. Performance comparison of different IQA models in four databases

      View table
      View in Article

      Table 2. Performance comparison of different IQA models in four databases

      Database CriteriaSSIMVIFMADFSIMcGMSDVSIGSMRVISMCVSSCAGSProposed
      TID2013SROCC0.741 70.676 90.780 70.851 00.804 40.896 50.794 60.675 70.806 90.831 60.859 2
      KROCC0.558 80.514 70.603 50.666 50.633 90.718 30.625 50.514 60.633 10.646 90.673 8
      PLCC0.789 50.772 00.826 70.876 90.858 30.900 00.846 40.782 50.840 60.844 50.878 1
      RMSE0.760 80.788 00.697 50.595 90.636 00.540 40.660 30.771 90.671 50.663 90.593 2
      TID2008SROCC0.774 90.749 10.834 00.884 00.890 70.897 90.850 40.737 50.896 10.823 10.891 6
      KROCC0.576 80.586 00.644 50.699 10.709 20.712 30.659 60.562 80.958 00.628 90.704 0
      PLCC0.773 20.808 40.830 80.876 20.878 80.876 20.842 20.795 40.595 60.809 10.886 8
      RMSE0.851 10.789 90.746 80.646 80.640 40.646 60.723 50.813 30.721 50.788 60.620 1
      CSIQSROCC0.875 60.919 50.946 60.931 00.957 00.942 30.910 80.897 90.958 00.919 80.950 5
      KROCC0.690 70.753 70.797 00.769 00.812 90.785 70.737 40.723 40.840 60.748 70.800 9
      PLCC0.861 30.927 70.950 20.919 20.954 10.927 90.896 40.923 60.958 90.901 40.953 7
      RMSE0.133 40.098 00.081 80.103 40.078 60.097 90.116 40.100 70.074 50.113 70.079 0
      LIVESROCC0.947 90.963 60.966 90.964 50.960 30.952 40.956 10.960 00.967 20.973 40.965 3
      KROCC0.796 30.828 20.842 10.836 30.826 80.805 80.815 00.820 30.840 60.865 80.835 6
      PLCC0.944 90.960 40.967 50.961 30.959 50.948 20.951 20.957 00.965 10.964 00.961 6
      RMSE8.945 57.613 76.907 37.529 67.693 78.681 68.432 77.927 47.157 38.325 17.498 3
      权重平均SROCC0.794 20.764 60.840 50.884 70.854 40.910 00.845 20.763 80.875 10.858 80.896 4
      KROCC0.610 80.604 90.670 20.710 10.702 20.736 60.673 20.600 60.711 50.682 80.730 0
      PLCC0.814 00.826 10.861 90.892 80.889 20.903 30.865 00.830 70.889 30.857 50.901 0
      直接平均SROCC0.835 00.827 30.882 10.907 60.903 10.922 30.878 00.817 80.908 00.887 00.916 7
      KROCC0.655 70.670 70.721 80.742 70.745 70.755 50.709 40.655 30.753 10.722 60.753 6
      PLCC0.842 20.867 10.893 80.908 40.912 70.913 10.884 10.864 60.915 20.879 80.920 1
    • Table 3. SROCC values of IQA model under different distortion types

      View table
      View in Article

      Table 3. SROCC values of IQA model under different distortion types

      DatabaseTypeSSIMVIFMADFSIMcGMSDVSIGSMRVISMCVSSCAGSProposed
      TID2013AGN0.867 10.899 40.884 30.910 10.946 20.946 00.906 40.864 50.940 10.935 90.932 5
      ANC0.772 60.829 90.801 90.853 70.868 40.870 50.817 50.812 40.863 90.865 30.843 4
      SCN0.851 50.883 50.891 10.890 00.935 00.936 70.915 80.841 20.907 70.927 60.911 9
      MN0.776 70.845 00.738 00.809 40.707 50.769 70.729 30.822 80.771 50.752 60.781 5
      HFN0.863 40.897 20.887 60.904 00.916 20.920 00.886 90.883 70.909 70.915 90.908 0
      IN0.750 30.853 70.276 90.825 10.763 70.874 10.796 50.882 30.745 70.836 10.781 4
      QN0.865 70.785 40.851 40.880 70.904 90.874 80.884 10.751 90.886 90.871 80.899 3
      GB0.966 80.965 00.931 90.955 10.911 30.961 20.968 90.974 20.934 80.961 40.926 8
      DEN0.925 40.891 10.925 20.933 00.952 50.948 40.943 20.894 70.942 70.946 60.946 6
      JPEG0.920 00.919 20.921 70.933 90.950 70.954 10.928 40.930 60.952 10.958 50.946 3
      JP2K0.946 80.951 60.951 10.958 90.965 70.970 60.960 20.952 10.958 70.962 00.954 7
      JGTE0.849 30.840 90.828 30.861 00.840 30.921 60.851 20.843 40.861 30.864 40.883 2
      J2TE0.882 80.876 10.878 80.891 90.913 60.922 80.918 20.885 40.885 10.925 00.880 5
      NEPN0.782 10.772 00.831 50.793 70.814 00.806 00.813 00.750 20.820 10.783 30.831 8
      Block0.572 00.530 60.281 20.553 20.662 50.171 30.641 80.603 70.515 20.601 50.662 0
      MS0.775 20.627 60.645 00.748 70.735 10.770 00.787 50.607 90.715 00.744 10.710 7
      CTC0.377 50.838 60.197 20.467 90.323 50.475 40.485 70.152 60.294 00.451 40.314 7
      CCS0.414 10.309 90.057 50.810 00.294 80.274 80.357 80.399 40.261 40.371 10.584 0
      MGN0.780 30.846 80.840 90.856 90.888 60.911 70.834 80.823 30.879 90.870 00.868 8
      CN0.856 60.894 60.906 40.924 30.929 80.912 10.912 40.898 40.935 10.916 80.934 5
      LCNI0.905 70.920 40.944 30.948 50.962 90.956 40.956 30.915 60.962 90.957 40.965 6
      ICQD0.854 20.841 40.874 50.881 50.910 20.883 90.897 30.803 60.910 80.906 00.906 1
      CHA0.877 50.884 80.831 00.892 50.853 00.890 60.882 30.915 40.852 30.876 80.857 7
      SSR0.946 10.935 30.956 70.957 60.968 30.962 80.966 80.943 90.960 50.958 00.960 3
      CSIQAGWN0.897 40.957 50.954 10.935 90.968 00.963 60.944 00.937 70.967 00.965 20.949 0
      JPEG0.954 60.970 50.961 50.966 40.965 00.961 80.963 20.959 90.968 90.957 30.968 2
      JP2K0.960 60.967 20.975 20.970 40.972 00.969 40.964 80.969 30.977 70.954 50.973 0
      AGPN0.892 20.951 10.957 00.937 00.950 00.963 80.938 70.931 90.951 60.949 20.932 6
      GB0.960 90.974 50.960 20.972 90.971 00.967 90.958 90.971 50.978 90.957 40.975 4
      GCD0.792 20.934 50.920 70.943 80.904 00.950 40.935 40.879 20.932 40.927 30.927 4
      LIVEJP2K0.961 40.969 60.967 60.972 40.971 00.960 40.970 00.962 20.971 90.982 20.970 7
      JPEG0.976 40.984 60.976 40.984 00.978 00.976 10.977 80.982 20.983 60.983 60.980 4
      AWGN0.969 40.985 80.984 40.971 60.974 00.983 50.977 40.969 20.980 90.983 70.966 9
      GB0.951 70.972 80.946 50.970 80.957 00.952 70.951 80.964 30.966 20.964 10.968 9
      FF0.955 60.965 00.956 90.951 90.942 00.943 00.940 20.958 10.959 20.963 30.968 0
    • Table 4. Comparison of efficiency of different IQA models

      View table
      View in Article

      Table 4. Comparison of efficiency of different IQA models

      IQA indexTime cost /sIQA indexTime cost/s
      SSIM0.278RVSIM0.939
      PSNR0.156MAD1.256
      CAGS0.483CPCCs0.531
      VSI0.536
    Tools

    Get Citation

    Copy Citation Text

    Benchi JIANG, Shilei BIAN, Chenyang SHI, Lulu WU. Full reference image quality assessment based on color appearance-based phase consistency[J]. Optics and Precision Engineering, 2023, 31(10): 1509

    Download Citation

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

    Category: Information Sciences

    Received: May. 14, 2022

    Accepted: --

    Published Online: Jul. 4, 2023

    The Author Email:

    DOI:10.37188/OPE.20233110.1509

    Topics