Table 2 Comparison of MAE, SSIM and SNRpeak of sCT produced by different methods in the whole pelvic region
From: Synthetic CT generation for pelvic cases based on deep learning in multi-center datasets
Method | MAE↓ | SNRpeak↑ | SSIM↑ | GP↑ |
|---|---|---|---|---|
CycleGAN | 47.129 | 25.167 | 0.978 | 0.545 |
CycleGAN (w/ CL loss) | 43.753 | 26.211 | 0.985 | 0.645 |
RTGAN | 43.704 | 25.344 | 0.982 | 0.66 |
2.5D RTGAN (w/ CL loss) | 42.557 | 26.268 | 0.986 | 0.690 |
2.5D CRGAN (w/ CL loss) | 42.344 | 26.979 | 0.992 | 0.911 |