• Acta Photonica Sinica
  • Vol. 51, Issue 4, 0410004 (2022)
Bendu BAI and Junpeng LI*
Author Affiliations
  • School of Communication and Information Engineering,Xi'an University of Posts & Telecommunications,Xi'an 710121,China
  • show less
    DOI: 10.3788/gzxb20225104.0410004 Cite this Article
    Bendu BAI, Junpeng LI. Multi-exposure Image Fusion Based on Attention Mechanism[J]. Acta Photonica Sinica, 2022, 51(4): 0410004 Copy Citation Text show less
    Network structure of AMEFNet
    Fig. 1. Network structure of AMEFNet
    Feature extraction network structure of AMEFNet
    Fig. 2. Feature extraction network structure of AMEFNet
    Attention network structure of AMEFNet
    Fig. 3. Attention network structure of AMEFNet
    Feature reconstruction network structure of Unet
    Fig. 4. Feature reconstruction network structure of Unet
    Loss function curve of AMEFNet
    Fig. 5. Loss function curve of AMEFNet
    Comparison of House sequence algorithm results
    Fig. 6. Comparison of House sequence algorithm results
    Comparison of TableLamp sequence algorithm results
    Fig. 7. Comparison of TableLamp sequence algorithm results
    Comparison of running time
    Fig. 8. Comparison of running time
    Exposure fusion results of Studio sequence
    Fig. 9. Exposure fusion results of Studio sequence
    Comparison of running time
    Fig. 10. Comparison of running time
    Fusion MethodPSNRAGSFENVIF
    GFF58.224 05.623 418.826 97.393 90.825 3
    DSIFT59.825 65.105 117.103 87.354 30.759 1
    SPD-MEF58.754 55.877 321.033 97.039 90.794 9
    MEFNet58.380 46.029 320.596 67.394 80.844 9
    MEF-GAN58.773 64.756 513.554 86.982 30.621 1
    Ours59.115 06.199 121.098 67.203 70.874 6
    Table 1. The values of five quality metrics averaged over the fused images on test set
    Fusion methodComplexityFLOPs(×109G)
    GFFONmn-
    DSIFTONmn-
    SPD-MEFONmn-
    MEFNet-13.2
    MEF-GAN-71.1
    Ours-165.6
    Table 2. Time complexity of the different algorithm
    Fusion methodPSNRAGSFENVIF
    RAMEFNet58.750 95.986 620.135 16.98190.833 9
    AMEFNet59.115 06.199 121.098 67.203 70.874 6
    Table 3. Evaluations of attention module on five quality metrics average
    Fusion methodPSNRAGSFENVIF
    AMEFNet-λ0.0259.249 86.035 120.113 47.10110.843 1
    AMEFNet-λ0.259.115 06.199 121.098 67.203 70.874 6
    AMEFNet-λ259.09906.174 321.035 97.19150.879 5
    Table 4. Evaluations of different hyperparameter λ on five quality metrics average
    Bendu BAI, Junpeng LI. Multi-exposure Image Fusion Based on Attention Mechanism[J]. Acta Photonica Sinica, 2022, 51(4): 0410004
    Download Citation