• Optics and Precision Engineering
  • Vol. 29, Issue 10, 2481 (2021)
Tao XU1,*, Ji-yong ZHOU2, Guo-liang ZHANG3, and Lei CAI1
Author Affiliations
  • 1School of Artificial Engineering, Henan Institute of Science and Technology, Xinxiang453003, China
  • 2School of Mechanical and Electrical Engineering, Henan Institute of Science and Technology, Xinxiang453003, China
  • 3Global Energy Interconnection Research Institute co.Ltd., Beijing102209, China
  • show less
    DOI: 10.37188/OPE.20212910.2481 Cite this Article
    Tao XU, Ji-yong ZHOU, Guo-liang ZHANG, Lei CAI. Fine restoration of incomplete image with external features and image features[J]. Optics and Precision Engineering, 2021, 29(10): 2481 Copy Citation Text show less
    Restoration results of the proposed restoration model in this paper
    Fig. 1. Restoration results of the proposed restoration model in this paper
    Schematic diagram of the proposed repair model in this paper
    Fig. 2. Schematic diagram of the proposed repair model in this paper
    Retrieval and acquisition of external features
    Fig. 3. Retrieval and acquisition of external features
    Network model of the proposed method for fine restoration of incomplete images
    Fig. 4. Network model of the proposed method for fine restoration of incomplete images
    Coherent schematic diagram of related features
    Fig. 5. Coherent schematic diagram of related features
    Comparison of the existing restoration model and the restoration model in this paper in the Place2 database, from right to left, the input images, SH, GLCI, CSA, roughly repaired image, EFIF, and the original image
    Fig. 6. Comparison of the existing restoration model and the restoration model in this paper in the Place2 database, from right to left, the input images, SH, GLCI, CSA, roughly repaired image, EFIF, and the original image
    Comparison of the existing restoration model and the restoration model in this paper in the RUIE database, from right to left, the input images, SH, GLCI, CSA, roughly repaired image, EFIF, and the original image
    Fig. 7. Comparison of the existing restoration model and the restoration model in this paper in the RUIE database, from right to left, the input images, SH, GLCI, CSA, roughly repaired image, EFIF, and the original image
    Comparison of the existing restoration model and the restoration model in this paper in the Underwater Target database, from right to left, the input images, SH, GLCI, CSA, roughly repaired image, EFIF, and the original image
    Fig. 8. Comparison of the existing restoration model and the restoration model in this paper in the Underwater Target database, from right to left, the input images, SH, GLCI, CSA, roughly repaired image, EFIF, and the original image
    Histograms of PSNR and SSIM values for restoration results from this restoration model and comparative restoration models in the Underwater Targe dataset.
    Fig. 9. Histograms of PSNR and SSIM values for restoration results from this restoration model and comparative restoration models in the Underwater Targe dataset.
    数据集掩码率PSNRSSIMLr(%)
    SHGLCICSA粗修复EFIFSHGLCLCSA粗修复EFIFSHGLCICSAEFIF
    Place210%-20%31.3431.2733.3631.0334.870.9060.9110.9890.9050.9871.021.000.780.71
    20%-30%30.2431.7932.8530.0833.120.8930.8970.9820.8940.9831.571.310.980.94
    30%-40%21.9624.1725.1621.3524.510.8860.8910.9260.8870.9312.742.361.291.23
    40%-50%20.3922.1824.5120.4024.840.8320.8530.8730.8290.8943.673.242.852.46
    RUIE10%-20%31.4533.7235.8431.5335.980.9290.9360.9530.9250.9781.551.471.181.06
    20%-30%30.7533.4634.9728.9535.730.9190.9250.9420.9170.9711.941.731.561.47
    30%-40%27.8427.7628.0627.4630.930.8970.9020.9160.8980.9232.952.862.722.55
    40%-50%21.9723.1923.9721.9924.450.8460.8740.8990.8410.9053.773.783.243.16
    Underwater Target10%-20%24.8325.0125.4224.6727.010.9130.9280.9360.9150.9491.651.441.111.09
    20%-30%24.3624.7225.3124.2126.750.9040.9160.9290.8990.9352.041.961.541.44
    30%-40%20.4821.1422.0720.8322.350.8840.8900.9070.8810.9163.272.982.732.49
    40%-50%15.3915.0416.4215.4616.910.8620.8760.8900.8680.8914.414.043.273.03
    Table 1. Objective data for restoration results in Place2, RUIE and Underwater Target database
    Tao XU, Ji-yong ZHOU, Guo-liang ZHANG, Lei CAI. Fine restoration of incomplete image with external features and image features[J]. Optics and Precision Engineering, 2021, 29(10): 2481
    Download Citation