Self-Reenactment on VOODOO-XP (test)
0.025MEt3RInstant Expressive Gaussian Head Avatar
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Instant Expressive Gaussian Head AvatarDimensionality=3D, Training Dataset=Original2025.12 | 0.025 | 0.8294 | 0.1864 | 0.496 | 107.31 | |
| InvertAvatarDimensionality=3D, Training Dataset=Original, Inference Optimization=Excluding morphable model fitting2025.12 | 0.026 | 0.8456 | 0.1592 | 0.564 | 24.3 | |
| GAGAvatar†Dimensionality=3D, Training Dataset=Synthetic, Inference Optimization=Excluding morphable model fitting2025.12 | 0.027 | 0.8258 | 0.1991 | 0.636 | 13.67 | |
| HYPortraitDimensionality=2D, Training Dataset=Original2025.12 | 0.028 | 0.8343 | 0.1685 | 0.483 | 0.01 | |
| VOODOO-XPDimensionality=3D, Training Dataset=Original2025.12 | 0.03 | 0.8086 | 0.1966 | 0.56 | 5.45 | |
| Portrait4D-v2Dimensionality=3D, Training Dataset=Original2025.12 | 0.03 | 0.8327 | 0.1709 | 0.545 | 14.5 | |
| Portrait4D-v2†Dimensionality=3D, Training Dataset=Synthetic2025.12 | 0.03 | 0.8217 | 0.1952 | 0.682 | 14.5 | |
| LivePortraitDimensionality=2D, Training Dataset=Original2025.12 | 0.032 | 0.8206 | 0.1739 | 0.4 | 78.12 | |
| X-NeMoDimensionality=2D, Training Dataset=Original2025.12 | 0.032 | 0.8227 | 0.174 | 0.416 | 0.03 | |
| GAGAvatarDimensionality=3D, Training Dataset=Original, Inference Optimization=Excluding morphable model fitting2025.12 | 0.032 | 0.8205 | 0.1914 | 0.501 | 13.67 |