• Laser & Optoelectronics Progress
  • Vol. 57, Issue 8, 081015 (2020)
Heng Li1、*, Liming Zhang2、3、**, Meirong Jiang2, and Yulong Li1
Author Affiliations
  • 1School of Electronics and Information Engineering, Lanzhou Jiaotong University, Lanzhou, Gansu 730070, China
  • 2Faculty of Geomatics, Lanzhou Jiaotong University, Lanzhou, Gansu 730070, China
  • 3National-Local Joint Engineering Research Center of Technologies and Applications for National Geographic State Monitoring, Lanzhou, Gansu 730070, China
  • show less
    DOI: 10.3788/LOP57.081015 Cite this Article Set citation alerts
    Heng Li, Liming Zhang, Meirong Jiang, Yulong Li. Multi-Focus Image Fusion Algorithm Based on Supervised Learning for Fully Convolutional Neural Networks[J]. Laser & Optoelectronics Progress, 2020, 57(8): 081015 Copy Citation Text show less
    Framework of proposed method
    Fig. 1. Framework of proposed method
    Network structure diagram
    Fig. 2. Network structure diagram
    Constructed data
    Fig. 3. Constructed data
    Training results of the 1st, 25th, 50th, 75th, and 100th epochs
    Fig. 4. Training results of the 1st, 25th, 50th, 75th, and 100th epochs
    Neural network loss function curve of training 100 epochs
    Fig. 5. Neural network loss function curve of training 100 epochs
    Partial of multi-focus test images used in experiment
    Fig. 6. Partial of multi-focus test images used in experiment
    Experimental results of first group images with different algorithms
    Fig. 7. Experimental results of first group images with different algorithms
    Experimental results of second group images with different algorithms
    Fig. 8. Experimental results of second group images with different algorithms
    IndexSSIMPSNRCCUQI
    P20.866354.08890.97530.7928
    P40.884257.14380.97930.8194
    P80.884257.60860.97930.8195
    P160.884257.60860.97930.8195
    C20.863153.56150.97430.7863
    C40.881356.84580.97880.8136
    C80.883757.67280.97950.8200
    C160.883757.67280.97950.8200
    Table 1. Average values of quality metrics for 27 test images under different fuzzy settings
    MethodSSIMPSNRCCUQI
    NSCT0.875857.10250.97780.8160
    DTCWT0.872456.91150.97690.8069
    CNN0.874256.94470.97650.8155
    NSST0.880657.98110.97930.8175
    Proposed0.883757.67280.97950.8200
    Table 2. Average values of quality metrics for 27 test images
    Heng Li, Liming Zhang, Meirong Jiang, Yulong Li. Multi-Focus Image Fusion Algorithm Based on Supervised Learning for Fully Convolutional Neural Networks[J]. Laser & Optoelectronics Progress, 2020, 57(8): 081015
    Download Citation