Surgical Video Generation on KASA NeedlePuncture (PSNR, SSIM, FID, FVD)
21.22PSNRKVLR
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| KVLRConditioning protocol=Visual-action-conditioned Generation2026.05 | 21.22 | 0.85 | 11.75 | 87.1 | |
| OpenSora-2.0Conditioning protocol=Visual-action-conditioned Generation2026.05 | 20.75 | 0.83 | 12.9 | 97.5 | |
| KVLR-fastConditioning protocol=Visual-action-conditioned Generation2026.05 | 20.55 | 0.81 | 14.6 | 112.5 | |
| SurgSoraConditioning protocol=Visual-action-conditioned Generation2026.05 | 19.25 | 0.76 | 19.35 | 140.6 | |
| Cosmos-H-SurgicalConditioning protocol=Raw-action-conditioned Generation2026.05 | 19.14 | 0.75 | 19.72 | 142.97 | |
| HunyuanVideoConditioning protocol=Raw-action-conditioned Generation2026.05 | 18.6 | 0.71 | 23.4 | 175.4 | |
| Wan-2.2Conditioning protocol=Raw-action-conditioned Generation2026.05 | 18.2 | 0.68 | 26.3 | 200.5 | |
| OpenSora-2.0Conditioning protocol=Raw-action-conditioned Generation2026.05 | 17.9 | 0.66 | 29.1 | 225.8 | |
| Cosmos-H-SurgicalConditioning protocol=Text-conditioned Generation2026.05 | 17.8 | 0.65 | 31.5 | 245.7 | |
| HunyuanVideoConditioning protocol=Text-conditioned Generation2026.05 | 17.1 | 0.6 | 37.2 | 285.2 | |
| Wan-2.2Conditioning protocol=Text-conditioned Generation2026.05 | 16.8 | 0.58 | 40.1 | 310.4 | |
| OpenSora-2.0Conditioning protocol=Text-conditioned Generation2026.05 | 16.6 | 0.56 | 42.3 | 335.6 |