• Laser & Optoelectronics Progress
  • Vol. 59, Issue 12, 1210005 (2022)
Peng Lu, Na Zhang, Guoliang Zou*, Zhenhua Wang, and Zongsheng Zheng
Author Affiliations
  • College of Information, Shanghai Ocean University, Shanghai 201306, China
  • show less
    DOI: 10.3788/LOP202259.1210005 Cite this Article Set citation alerts
    Peng Lu, Na Zhang, Guoliang Zou, Zhenhua Wang, Zongsheng Zheng. CycleGAN Coastline Automatic Extraction Method Based on Dual Attention Mechanism[J]. Laser & Optoelectronics Progress, 2022, 59(12): 1210005 Copy Citation Text show less
    CycleGAN architecture
    Fig. 1. CycleGAN architecture
    DAM-CycleGAN algorithm flowchart
    Fig. 2. DAM-CycleGAN algorithm flowchart
    CycleGAN network architecture based on dual attention mechanism
    Fig. 3. CycleGAN network architecture based on dual attention mechanism
    Generator_DA model structure based on dual attention mechanism
    Fig. 4. Generator_DA model structure based on dual attention mechanism
    Experimental data. (a) Original dataset; (b) dataset used in this paper; (c) Google Maps converted into sea-land binarized image by sea-land binarization operation
    Fig. 5. Experimental data. (a) Original dataset; (b) dataset used in this paper; (c) Google Maps converted into sea-land binarized image by sea-land binarization operation
    Converted sea-land binarization effect maps (100 pairs of training set)
    Fig. 6. Converted sea-land binarization effect maps (100 pairs of training set)
    Experimental results of different models (300 pairs of training sets)
    Fig. 7. Experimental results of different models (300 pairs of training sets)
    MethodMSEMPAMIoU
    Method one2648.660.82190.8956
    Method two2052.870.89530.9179
    Method three2655.680.89220.9074
    Method four2047.310.91000.9241
    Method five1642.980.92010.9363
    Table 1. Comparison results of MSE, MPA, and MIoU on 35 pairs of training sets and 50 pairs of test sets
    MethodMSEMPAMIoU
    Method one4207.070.82080.8389
    Method two1857.450.90130.9157
    Method three1323.260.93050.9384
    Method four1400.980.91780.9268
    Method five1126.540.92580.9390
    Table 2. Comparison results of MSE, MPA, and MIoU on 100 pairs of training sets and 50 pairs of test sets
    ModelMSEMPAMIoU
    FCN3685.870.93840.8608
    DeepLab3761.870.93510.8736
    DAM-CycleGAN1043.040.93100.9533
    Table 3. Comparison results of MSE, MPA, and MIoU on 200 pairs of training sets and 30 pairs of test sets
    ModelMSEMPAMIoU
    FCN4533.150.92590.8852
    DeepLab3125.880.94530.8944
    DAM-CycleGAN946.400.93350.9568
    Table 4. Comparison results of MSE, MPA, and MIoU on 300 pairs of training sets and 30 pairs of test sets
    Peng Lu, Na Zhang, Guoliang Zou, Zhenhua Wang, Zongsheng Zheng. CycleGAN Coastline Automatic Extraction Method Based on Dual Attention Mechanism[J]. Laser & Optoelectronics Progress, 2022, 59(12): 1210005
    Download Citation