Medical Image Synthesis on VoCo 10k (train/test)
0.7874FIDMeDUET
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MeDUETBackbone=SiT-B/4, Pretraining steps=200k, Conditioning=learned content and style vectors2026.02 | 0.7874 | 0.5598 | 0.1659 | |
| MeDUETBackbone=SiT-B/4, Pretraining steps=100k, Conditioning=learned content and style vectors2026.02 | 0.7908 | 0.5627 | 0.1677 | |
| MeDUETBackbone=SiT-B/4, Pretraining steps=200k, Conditioning=pre-defined metadata vectors2026.02 | 0.8011 | 0.5782 | 0.1692 | |
| MeDUETBackbone=SiT-B/4, Pretraining steps=100k, Conditioning=pre-defined metadata vectors2026.02 | 0.8039 | 0.5806 | 0.1712 | |
| SiT-B/4Conditioning=learned content and style vectors2026.02 | 0.8533 | 0.6012 | 0.1798 | |
| MeDUETBackbone=DiT-B/4, Pretraining steps=200k, Conditioning=learned content and style vectors2026.02 | 0.8611 | 0.6003 | 0.1803 | |
| MeDUETBackbone=DiT-B/4, Pretraining steps=100k, Conditioning=learned content and style vectors2026.02 | 0.8642 | 0.6028 | 0.1824 | |
| SiT-B/4Conditioning=pre-defined metadata vectors2026.02 | 0.867 | 0.6023 | 0.1834 | |
| MeDUETBackbone=DiT-B/4, Pretraining steps=200k, Conditioning=pre-defined metadata vectors2026.02 | 0.8727 | 0.6074 | 0.1876 | |
| MeDUETBackbone=DiT-B/4, Pretraining steps=100k, Conditioning=pre-defined metadata vectors2026.02 | 0.8763 | 0.6097 | 0.1892 | |
| DiT-B/4Conditioning=learned content and style vectors2026.02 | 0.9074 | 0.6175 | 0.1906 | |
| MAISIModel Category=Medical Image Synthesis Models2026.02 | 0.9139 | 0.6292 | 0.2057 | |
| DiT-B/4Conditioning=pre-defined metadata vectors2026.02 | 0.9207 | 0.6329 | 0.1957 | |
| 3D MedDiffusionModel Category=Medical Image Synthesis Models2026.02 | 0.9216 | 0.6327 | 0.2032 | |
| WDMModel Category=Medical Image Synthesis Models2026.02 | 0.9668 | 0.6612 | 0.2284 | |
| MedSynModel Category=Medical Image Synthesis Models2026.02 | 0.9873 | 0.6734 | 0.2325 |