notesum.ai
Published at October 23VISAGE: Video Synthesis using Action Graphs for Surgery
cs.CL
cs.AI
cs.LG
Released Date: October 23, 2024
Authors: Yousef Yeganeh1, Rachmadio Lazuardi1, Amir Shamseddin1, Emine Dari1, Yash Thirani1, Nassir Navab Azade Farshad
Aff.: 1Technical University of Munich

| Model | Input | FVD (↓) | PSNR (↑) | LPIPS (↓) | SSIM (↑) |
|---|---|---|---|---|---|
| CoDi [34] | Image + Triplet | 6,944 | 9.8 | 0.82 | 0.31 |
| WALDO [24] | Image + Seg. + Flow | 3,413 | 11.6 | 0.72 | 0.34 |
| LFDM [26] | Image + Triplet | 1,957 | 12.0 | 0.54 | 0.71 |
| SVD [4] | Image | 3,870 | 14.8 | 0.51 | 0.47 |
| SVD + FT [4] | Image | 1,931 | 18.2 | 0.40 | 0.55 |
| VISAGE-T (Ours) | Image + Triplet | 1,780 | 18.1 | 0.39 | 0.56 |
| VISAGE-I (Ours) | Image + Triplet | 1,875 | 18.3 | 0.38 | 0.56 |