• Optics and Precision Engineering
  • Vol. 31, Issue 16, 2406 (2023)
Baoping LI, Hengyi QI*, Manli WANG, and Po WEI
Author Affiliations
  • College of Physics and Electronic Information, Henan Polytechnic University, Jiaozuo454000, China
  • show less
    DOI: 10.37188/OPE.20233116.2406 Cite this Article
    Baoping LI, Hengyi QI, Manli WANG, Po WEI. Equipment fault dataset amplification method combine 3D model with improved CycleGAN[J]. Optics and Precision Engineering, 2023, 31(16): 2406 Copy Citation Text show less
    Framework of CycleGAN
    Fig. 1. Framework of CycleGAN
    Framework of improved generator
    Fig. 2. Framework of improved generator
    Feature map of improved generator
    Fig. 3. Feature map of improved generator
    Several generator simplification architectures
    Fig. 4. Several generator simplification architectures
    Contour feature images
    Fig. 5. Contour feature images
    Framework of discriminator
    Fig. 6. Framework of discriminator
    Loss function
    Fig. 7. Loss function
    Framework of 3D modeling
    Fig. 8. Framework of 3D modeling
    Camera perspective
    Fig. 9. Camera perspective
    Fake Y for multiple network outputs
    Fig. 10. Fake Y for multiple network outputs
    MS_SSIM metrics for three generators
    Fig. 11. MS_SSIM metrics for three generators
    Loss curve for three generators
    Fig. 12. Loss curve for three generators
    Generator/GANIS ⬆FID ⬇PSNR ⬆
    DCGAN1.385 7178.805 5/
    WGAN1.375 3161.486 5/
    WGAN-GP1.417 4155.687 1/
    StyleGAN1.397 6164.547 3/
    ChipGAN1.364 1287.971 226.467 4
    ResNet1.484 9232.266 934.595 4
    U-net1.446 9227.299 133.245 7
    Ours1.519 7192.021 235.096 4
    Table 1. Result of three evaluation indicators
    数据集左偏AP右偏APMAP
    真实86.795.493.6
    WGAN-GP+真实89.896.194.8
    U-net+真实91.898.695.7
    ResNet+真实93.399.196.2
    Ours+真实96.799.598.1
    Table 2. Results of Yolov5 detection
    ModelParams size /MBLayers of GGPU Memory Usage/GB
    ResNet43.42925.0
    U-net207.57544.6
    U-ResNet41.7755.8
    Table 3. Generator parameters and training memory occupancy
    Baoping LI, Hengyi QI, Manli WANG, Po WEI. Equipment fault dataset amplification method combine 3D model with improved CycleGAN[J]. Optics and Precision Engineering, 2023, 31(16): 2406
    Download Citation