3D Camera-Controlled Video Synthesis on MSR-VTT unseen camera trajectories
0.428TransErrAC3D
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| AC3DArchitecture=VDiT2024.11 | 0.428 | 0.038 | 5.34 | 110.71 | 28.58 | |
| AC3DArchitecture=CogVideoX2024.11 | 0.431 | 0.039 | 5.52 | 116.04 | 28.38 | |
| VD3DArchitecture=VDiT2024.11 | 0.486 | 0.047 | 6.88 | 137.62 | 27.9 | |
| VD3DArchitecture=CogVideoX2024.11 | 0.501 | 0.068 | 7.45 | 148.11 | 27.65 | |
| VD3DArchitecture=FIT2024.11 | 0.504 | 0.05 | 7.8 | 165.18 | 26.89 | |
| CameraCtrlArchitecture=VDiT2024.11 | 0.566 | 0.143 | 8.15 | 146.77 | 27.51 | |
| CameraCtrlArchitecture=U-Net2024.11 | 0.587 | 0.132 | 12.33 | 201.33 | 25.05 | |
| MotionCtrlArchitecture=VDiT2024.11 | 0.589 | 0.146 | 9.92 | 150.2 | 27.25 | |
| MotionCtrlArchitecture=U-Net2024.11 | 0.593 | 0.137 | 16.85 | 283.12 | 24.11 |