• Laser & Optoelectronics Progress
  • Vol. 56, Issue 7, 071102 (2019)
Liu Zhuo1,2, Xiaoqi Chen1,2, Zhenping Xie1,2,*, Xiaojun Jiang3, and Daokun Bi3
Author Affiliations
  • 1 School of Digital Media, Jiangnan University, Wuxi, Jiangsu 214122, China
  • 2 Jiangsu Key Laboratory of Media Design and Software Technology, Wuxi, Jiangsu 214122, China
  • 3 Science and Technology on Near-Surface Detection Laboratory, Wuxi, Jiangsu 214035, China
  • show less
    DOI: 10.3788/LOP56.071102 Cite this Article Set citation alerts
    Liu Zhuo, Xiaoqi Chen, Zhenping Xie, Xiaojun Jiang, Daokun Bi. Simulation Learning Method for Discovery of Camouflage Targets Based on Deep Neural Networks[J]. Laser & Optoelectronics Progress, 2019, 56(7): 071102 Copy Citation Text show less
    Simulation learning framework for discovery of camouflage targets
    Fig. 1. Simulation learning framework for discovery of camouflage targets
    Simulation model of camouflage scenes
    Fig. 2. Simulation model of camouflage scenes
    Simulation results of camouflage scenes
    Fig. 3. Simulation results of camouflage scenes
    Preprocessing of scene area in which target can be placed. (a) Original scene map; (b) preprocessed binary map
    Fig. 4. Preprocessing of scene area in which target can be placed. (a) Original scene map; (b) preprocessed binary map
    Parallel dilated convolution
    Fig. 5. Parallel dilated convolution
    Multi-scale discovery model of camouflage targets
    Fig. 6. Multi-scale discovery model of camouflage targets
    Experimental discovery output results of representative camouflage targets. (a)(c) Test input images; (b) (d) corresponding model test outputs
    Fig. 7. Experimental discovery output results of representative camouflage targets. (a)(c) Test input images; (b) (d) corresponding model test outputs
    MethodPARMIOU
    Scale1(0.5)93.1790.3588.77
    Scale2(0.75)94.6891.0588.84
    Scale3(1.0)96.1591.9389.14
    Scale1+296.4292.1489.42
    Out(Scale1+2+3)98.7893.5690.96
    Table 1. Performance results obtained by proposed method at different scale parameters%
    MethodPARMIOU
    FCN86.1785.3576.71
    Pix2Pix93.6889.3586.88
    Out98.7893.5690.96
    Table 2. Discovery performance results of camouflage targets obtained by different methods%
    Expert interpretation levelPARMIOU
    Great78.0485.0171.60
    Good91.6890.7484.78
    Bad93.9991.5487.12
    Table 3. Model performance results obtained on a group of real scene images%
    Liu Zhuo, Xiaoqi Chen, Zhenping Xie, Xiaojun Jiang, Daokun Bi. Simulation Learning Method for Discovery of Camouflage Targets Based on Deep Neural Networks[J]. Laser & Optoelectronics Progress, 2019, 56(7): 071102
    Download Citation