Fig 1.
The schematic diagram of the generative adversarial networks.
Fig 2.
The structure of the self-attention.
Fig 3.
The structure of discriminative network.
Table 1.
Performance indicators of images under different hyperparameters.
Fig 4.
(a) The loss function in training process; (b) The training and validation mean square error.
Fig 5.
The eighteen defect template samples: (a) twelve training template samples; (b) three testing template samples; (c) three validation samples.
Table 2.
Comparison of PSNR and SSIM between the proposed method and baseline methods.
Fig 6.
Experimental template.
Fig 7.
Positron reconstruction images.
Table 3.
Comparison of PSNR and SSIM values for reconstructed images.
Fig 8.
The related parameters of the hydraulic cylinder: material is alloy steel;hydraulic oil is a water glycol flame-retardant hydraulic fluid HOUGHTO-SAFE 620C; outer diameter is 55 mm; inside diameter is 45 mm; wall thickness is 5 mm.
Fig 9.
SolidWorks simulation model.
Fig 10.
Experimental parameters: the concentration of nuclide is 800 bq; the sampling time is 10 s; the material is the iron wire (foreign body) in the cavity.
Table 4.
Comparison of PSNR and SSIM values for reconstructed images.