Lung CT Synthesis on LIDC-IDRI
0.0462MMD (Full Range)multi-head VQVAE
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| multi-head VQVAEApproach=Multi-head, Reconstruction Model=MLP02026.03 | 0.0462 | 0.0656 | 0.0479 | 0.0283 | 0.0212 | |
| Score-based DMApproach=Baseline, Reconstruction Model=-2026.03 | 0.0481 | 0.0531 | 0.0269 | 0.0342 | 0.0407 | |
| multi-head VQVAEApproach=Multi-head, Reconstruction Model=CNN3×3×32026.03 | 0.0517 | 0.0461 | 0.0307 | 0.0221 | 0.0209 | |
| multi-decoder VQVAEApproach=Multi-decoder, Reconstruction Model=CNN3×3×32026.03 | 0.052 | 0.0489 | 0.0485 | 0.0297 | 0.0247 | |
| multi-decoder VQVAEApproach=Multi-decoder, Reconstruction Model=MLP02026.03 | 0.055 | 0.0684 | 0.0493 | 0.0397 | 0.0214 | |
| VQVAEApproach=Baseline, Reconstruction Model=-2026.03 | 0.0631 | 0.0891 | 0.0454 | 0.0728 | 0.0824 | |
| multi-channel VQVAEApproach=Multi-channel, Reconstruction Model=MLP02026.03 | 0.0656 | 0.0551 | 0.0292 | 0.0384 | 0.0203 | |
| multi-channel VQVAEApproach=Multi-channel, Reconstruction Model=CNN3×3×32026.03 | 0.0676 | 0.049 | 0.0427 | 0.0335 | 0.0354 | |
| Score-based DMApproach=Multi-channel, Reconstruction Model=CNN3×3×32026.03 | 0.0795 | 0.0955 | 0.1754 | 0.1326 | 0.2076 | |
| Score-based DMApproach=Multi-channel, Reconstruction Model=MLP02026.03 | 0.0984 | 0.0969 | 0.1772 | 0.1298 | 0.2053 | |
| WGAN-GPApproach=Baseline, Reconstruction Model=-2026.03 | 0.17 | 0.0893 | 0.1451 | 0.1578 | 0.1624 | |
| WGAN-GPApproach=Multi-channel, Reconstruction Model=CNN3×3×32026.03 | 0.2314 | 0.1184 | 0.1319 | 0.0994 | 0.1869 | |
| WGAN-GPApproach=Multi-channel, Reconstruction Model=MLP02026.03 | 0.2806 | 0.1996 | 0.1508 | 0.09 | 0.2056 |