Latent Prediction on CLEVRER
0.982D Latent L2 LossNeural Voxel Dynamics (GT)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Neural Voxel Dynamics (GT)Camera setup=Single-camera, Depth source=GT depth2026.06 | 0.98 | 0.02 | |
| Neural Voxel Dynamics (GT)Camera setup=Single-camera, Depth source=estimated depth2026.06 | 1.01 | 0.39 | |
| Neural Voxel Dynamics (estimate)Camera setup=Single-camera, Depth source=estimated depth2026.06 | 1.01 | 0.36 | |
| Neural Voxel Dynamics (GT)Camera setup=Multi-camera, Depth source=GT depth2026.06 | 1.12 | 0.82 | |
| Neural Voxel Dynamics (estimate)Camera setup=Single-camera, Depth source=GT depth2026.06 | 1.26 | 0.25 | |
| 2D baselineCamera setup=Single-camera, Depth source=GT depth2026.06 | 1.39 | 0.22 | |
| 2D baselineCamera setup=Single-camera, Depth source=estimated depth2026.06 | 1.53 | 0.42 | |
| PhysGenCamera setup=Single-camera, Depth source=GT depth2026.06 | 1.91 | 0.17 | |
| Neural Voxel Dynamics (estimate)Camera setup=Multi-camera, Depth source=estimated depth2026.06 | 2.25 | 0.9 | |
| Neural Voxel Dynamics (estimate)Camera setup=Multi-camera, Depth source=GT depth2026.06 | 2.5 | 0.99 | |
| Neural Voxel Dynamics (GT)Camera setup=Multi-camera, Depth source=estimated depth2026.06 | 2.51 | 0.91 | |
| PhysCtrlCamera setup=Single-camera, Depth source=estimated depth2026.06 | 2.84 | 0.57 | |
| PhysCtrlCamera setup=Single-camera, Depth source=GT depth2026.06 | 2.96 | 0.35 | |
| PhysGenCamera setup=Single-camera, Depth source=estimated depth2026.06 | 2.97 | 0.56 | |
| CogVideoXCamera setup=Single-camera, Depth source=GT depth2026.06 | 2.98 | 0.32 | |
| CogVideoXCamera setup=Single-camera, Depth source=estimated depth2026.06 | 2.98 | 0.56 | |
| 2D baselineCamera setup=Multi-camera, Depth source=GT depth2026.06 | 3.19 | 1.2 | |
| PhysCtrlCamera setup=Multi-camera, Depth source=estimated depth2026.06 | 3.21 | 1.15 | |
| PhysCtrlCamera setup=Multi-camera, Depth source=GT depth2026.06 | 3.32 | 1.01 | |
| PhysGenCamera setup=Multi-camera, Depth source=GT depth2026.06 | 3.57 | 1.89 | |
| 2D baselineCamera setup=Multi-camera, Depth source=estimated depth2026.06 | 3.69 | 1.1 | |
| CogVideoXCamera setup=Multi-camera, Depth source=GT depth2026.06 | 3.87 | 1.5 | |
| PhysGenCamera setup=Multi-camera, Depth source=estimated depth2026.06 | 4.12 | 1.24 | |
| PhysGaussianCamera setup=Single-camera, Depth source=GT depth2026.06 | 4.13 | 0.56 | |
| CogVideoXCamera setup=Multi-camera, Depth source=estimated depth2026.06 | 4.14 | 1.18 | |
| PhysGaussianCamera setup=Multi-camera, Depth source=estimated depth2026.06 | 4.21 | 1.14 | |
| PhysGaussianCamera setup=Multi-camera, Depth source=GT depth2026.06 | 4.34 | 1.18 | |
| PhysGaussianCamera setup=Single-camera, Depth source=estimated depth2026.06 | 4.39 | 0.55 |