Research map / Video / Video generation
LooseControlVideo: Directorial Video Control using Spatial Blocking
Shariq Farooq Bhat, Kalyan Sunkavalli, Niloy Mitra
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace 2D box control with rendered oriented 3D boxes (DNOCS) (ablation) | VACE ft on Wan 2.2 · nuScenes val | TrajErr: 16.66 -> 5.79 | Improved |
| replace 2D box control with rendered oriented 3D boxes (DNOCS) (ablation) | VACE ft on Wan 2.2 · nuScenes val | OcclAcc: 42.45 -> 92.69 | Improved |
| replace 2D box control with rendered oriented 3D boxes (DNOCS) (ablation) | VACE ft on Wan 2.2 · nuScenes val | Contain: 96.33 -> 87.93 | Hurt |
| replace 2D optical flow control with rendered oriented 3D boxes (ablation) | VACE ft on Wan 2.2 · nuScenes val | RMC: 0.528 -> 0.318 | Improved |
| replace 2D optical flow control with rendered oriented 3D boxes (ablation) | VACE ft on Wan 2.2 · nuScenes val | OcclAcc: 79.32 -> 92.69 | Improved |
| fine-tune VACE on the authors' training set with 2D flow control | VACE (2D Flow) · nuScenes val | TrajErr: 7.86 -> 6.78 | Improved |
| add control signal (2D flow) versus control-free first+last frame baseline | VACE · nuScenes val | TrajErr: 90.12 -> 7.86 | Improved |
| replace 2D optical flow control with rendered oriented 3D boxes (ablation) | VACE ft on Wan 2.2 · HO-3D | TrajErr: 5.4 -> 3.9 | Improved |
| replace 2D box control with rendered oriented 3D boxes (ablation) | VACE ft on Wan 2.2 · HO-3D | OcclAcc: 55.1 -> 94.1 | Improved |
| replace 2D box control with rendered oriented 3D boxes (ablation) | VACE ft on Wan 2.2 · BEHAVE | TrajErr: 14.9 -> 5.8 | Improved |
| replace 2D optical flow control with rendered oriented 3D boxes (ablation) | VACE ft on Wan 2.2 · BEHAVE | OcclAcc: 78.5 -> 90.2 | Improved |
| use rendered oriented 3D boxes instead of 2D boxes (ablation) | VACE ft on Wan 2.2 · HO-3D | Contain: 97.9 -> 91.3 | Hurt |
Closest papers by meaning
- 3D-Aware Implicit Motion Control for View-Adaptive Human Video Generation (CVPR 2026 · similarity 0.90)
- Wan-Move: Motion-controllable Video Generation via Latent Trajectory Guidance (NeurIPS 2025 · similarity 0.89)
- GEN3C: 3D-Informed World-Consistent Video Generation with Precise Camera Control (CVPR 2025 · similarity 0.88)
- Training-free Camera Control for Video Generation (ICLR 2025 · similarity 0.88)
- 3DTrajMaster: Mastering 3D Trajectory for Multi-Entity Motion in Video Generation (ICLR 2025 · similarity 0.87)
- ONE-SHOT: Compositional Human-Environment Video Synthesis via Spatial-Decoupled Motion Injection and Hybrid Context Integration (NeurIPS 2026 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
