• Laser & Optoelectronics Progress
  • Vol. 59, Issue 4, 0410010 (2022)
Guoyang Chen1、2, Xiaojun Wu1、2、*, and Tianyang Xu2
Author Affiliations
  • 1School of Artificial Intelligence and Computer Science, Jiangnan University, Wuxi , Jiangsu 214122, China
  • 2Jiangsu Provincial Engineering Laboratory of Pattern Recognition and Computational Intelligence, Jiangnan University, Wuxi , Jiangsu 214122, China
  • show less
    DOI: 10.3788/LOP202259.0410010 Cite this Article Set citation alerts
    Guoyang Chen, Xiaojun Wu, Tianyang Xu. Unsupervised Infrared Image and Visible Image Fusion Algorithm Based on Deep Learning[J]. Laser & Optoelectronics Progress, 2022, 59(4): 0410010 Copy Citation Text show less
    Structure of proposed network
    Fig. 1. Structure of proposed network
    Diagram of fusion layer
    Fig. 2. Diagram of fusion layer
    Experiments on "street" images. (a) Infrared image; (b) visible light image; (c) fusion result of experiment 1; (d) fusion result of experiment 2; (e) fusion result of experiment 3
    Fig. 3. Experiments on "street" images. (a) Infrared image; (b) visible light image; (c) fusion result of experiment 1; (d) fusion result of experiment 2; (e) fusion result of experiment 3
    Experiments on "man" images. (a) Infrared image; (b) visible light image; (c) fusion result of experiment 1; (d) fusion result of experiment 2; (e) fusion result of experiment 3
    Fig. 4. Experiments on "man" images. (a) Infrared image; (b) visible light image; (c) fusion result of experiment 1; (d) fusion result of experiment 2; (e) fusion result of experiment 3
    Experiments on 3 pairs of images. (a) Infrared images; (b) visible light images; (c) fusion results of experiment 1; (d) fusion results of experiment 2; (e) fusion results of experiment 3
    Fig. 5. Experiments on 3 pairs of images. (a) Infrared images; (b) visible light images; (c) fusion results of experiment 1; (d) fusion results of experiment 2; (e) fusion results of experiment 3
    Experiments with different algorithms on "street" images
    Fig. 6. Experiments with different algorithms on "street" images
    Experiments with different algorithms on "man" images
    Fig. 7. Experiments with different algorithms on "man" images
    Experiments with different algorithms on 6 pairs of images
    Fig. 8. Experiments with different algorithms on 6 pairs of images
    ExperimentENSDVIFMIQabfSCD
    Experiment 16.594963.70260.610713.18980.57641.7738
    Experiment 26.719071.70950.648313.43790.56831.7631
    Experiment 36.796979.65000.673913.59380.60231.8067
    Table 1. Average quality metrics of 21 fused images in different experiments
    AlgorithmENSDVIFMIQabfSCD
    DCHWT6.567864.97890.505613.13550.46591.6099
    JSR6.364474.11590.640712.72870.35931.7517
    GTF6.635367.62600.413613.27070.42471.0134
    WLS6.640770.58890.728713.28140.50081.7961
    CSR6.258750.74370.392212.51740.53481.6482
    VGGML6.181948.13850.294912.36390.36771.6348
    DenseFuse6.724766.00130.660113.44930.40091.8083
    FusionGAN6.362954.35750.453512.72570.21831.4569
    IFCNN6.595566.87580.590313.19090.50331.7137
    Proposed algorithm6.796979.65000.673913.59380.60231.8067
    Table 2. Average index of 21 fused images
    Guoyang Chen, Xiaojun Wu, Tianyang Xu. Unsupervised Infrared Image and Visible Image Fusion Algorithm Based on Deep Learning[J]. Laser & Optoelectronics Progress, 2022, 59(4): 0410010
    Download Citation