• Acta Optica Sinica
  • Vol. 39, Issue 3, 0311002 (2019)
Jiangrong Xie1、2, Fanming Li1、3、*, Hong Wei1, and Bing Li1
Author Affiliations
  • 1 Shanghai Institute of Technical Physics, Chinese Academy of Sciences, Shanghai 200083, China
  • 2 University of Chinese Academy of Sciences, Beijing 100049, China
  • 3 Key Laboratory of Infrared System Detection and Imaging Technology, Chinese Academy of Sciences, Shanghai 200083, China
  • show less
    DOI: 10.3788/AOS201939.0311002 Cite this Article Set citation alerts
    Jiangrong Xie, Fanming Li, Hong Wei, Bing Li. Infrared Target Simulation Method Based on Generative Adversarial Neural Networks[J]. Acta Optica Sinica, 2019, 39(3): 0311002 Copy Citation Text show less
    Framework of CGAN model
    Fig. 1. Framework of CGAN model
    Structural diagram of generation network of DCGAN
    Fig. 2. Structural diagram of generation network of DCGAN
    Flow chart of simulation algorithm based on C-DCGAN model
    Fig. 3. Flow chart of simulation algorithm based on C-DCGAN model
    Structural diagram of generation network of C-DCGAN
    Fig. 4. Structural diagram of generation network of C-DCGAN
    Structural diagram of discrimination network of C-DCGAN
    Fig. 5. Structural diagram of discrimination network of C-DCGAN
    Trends of loss function on infrared dataset. (a) Ld_loss_real; (b) Ld_loss_fake; (c) Ld_loss; (d) Lg_loss
    Fig. 6. Trends of loss function on infrared dataset. (a) Ld_loss_real; (b) Ld_loss_fake; (c) Ld_loss; (d) Lg_loss
    Effect image of infrared targets generated by C-DCGAN
    Fig. 7. Effect image of infrared targets generated by C-DCGAN
    Effect image of MNIST samples generated by C-DCGAN
    Fig. 8. Effect image of MNIST samples generated by C-DCGAN
    MethodMNISTIR dataset
    Linear classifier (1-layer NN)91.6080.16
    K-nearest-neighbors, Euclidean (L2)95.0081.44
    C-DCGAN+L2 SVM93.6984.31
    Table 1. Accuracy comparison among different classification methods%
    MethodMNISTIR dataset
    Copy95.4082.22
    Affine transformation96.8384.17
    C-DCGAN generator96.1585.61
    Table 2. Performance comparison among different data augmentation methods%
    Jiangrong Xie, Fanming Li, Hong Wei, Bing Li. Infrared Target Simulation Method Based on Generative Adversarial Neural Networks[J]. Acta Optica Sinica, 2019, 39(3): 0311002
    Download Citation