Table 1 Hyperparameter settings for OctSinGAN training based on different TIs.
Hyperparameter names | TI categories | ||
|---|---|---|---|
a | b | c | |
\(\:min\_size\) (unit: voxel) | 10×10×10 | 20×20×20 | 10×10×10 |
\(\:num\_train\) | 5 | 5 | 5 |
\(\:{Oct}_{layer}\) | 5 | 5 | 5 |
\(\:lr\) | 0.0001 | 0.0001 | 0.0001 |
\(\:{\lambda\:}_{gp}\) | 0.1 | 0.1 | 0.1 |
\(\:\alpha\:\) | 5 | 10 | 10 |
\(\:epoch\) | 600 | 600 | 600 |
Optimizer category | Adam with default settings (Kingma and Ba, 2014) | ||
