Research map / Point clouds / Point cloud classification
Vista4D: Video Reshooting with 4D Point Clouds
Kuan Heng Lin, Zhizheng Liu, Pablo Salamanca, Yash Kant, Ryan Burgert, Yuancheng Xu, Koichi Namekata, Yiwei Zhao, Bolei Zhou, Micah Goldblum, Paul Debevec, Ning Yu
Read the paperCheck your idea against 137,217 papers
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 |
|---|---|---|---|
| add training with depth artifacts and in-context source video conditioning (ablation) | Vista4D · robustness to 4D reconstruction artifacts | qualitative | Improved |
| add temporal persistence of the 4D point cloud (ablation) | Vista4D · low-overlap source video and target cameras | static content preservation and camera accuracy | Improved |
| use point-cloud conditioning with 4D grounding vs implicit-prior ReCamMaster | Vista4D · camera control | translation error: 1.574 -> 1.251 | Improved |
| use Vista4D instead of GEN3C | Vista4D · 3D consistency | RE@SG: 12.99 -> 7.504 | Improved |
| use Vista4D instead of GEN3C | Vista4D · iPhone novel-view synthesis | EPE: 3.019 -> 1.142 | Improved |
| use Vista4D instead of TrajectoryCrafter | Vista4D · iPhone novel-view synthesis | SSIM: 0.320 -> 0.310 | Hurt |
| use Vista4D instead of ReCamMaster | Vista4D · video fidelity | FID: 94.15 -> 105.4 | Hurt |
| use Vista4D instead of GEN3C | Vista4D · video fidelity | FID: 113.5 -> 105.4 | Improved |
| use Vista4D instead of TrajectoryCrafter | Vista4D · VBench human anatomy | human anatomy: 0.790 -> 0.857 | Improved |
| use Vista4D instead of GEN3C | Vista4D · user study overall fidelity | preference: 5.952% -> 77.38% | Improved |
| use Vista4D 1280x720 instead of 672x384 | Vista4D · inference time | model inference (s): 1195 -> 9924 | Hurt |
| use segmentation failure (racket not segmented) | Vista4D · qualitative | streaking artifacts | No clear effect |
Closest papers by meaning
- 4K4D: Real-Time 4D View Synthesis at 4K Resolution (CVPR 2024 · similarity 0.87)
- Restage4D: Reanimating Deformable 3D Reconstruction from a Single Video (NeurIPS 2025 · similarity 0.87)
- CAT4D: Create Anything in 4D with Multi-View Video Diffusion Models (CVPR 2025 · similarity 0.86)
- X4D-SceneFormer: Enhanced Scene Understanding on 4D Point Cloud Videos through Cross-Modal Knowledge Transfer (AAAI 2024 · similarity 0.85)
- Instant4D: 4D Gaussian Splatting in Minutes (NeurIPS 2025 · similarity 0.85)
- EG4D: Explicit Generation of 4D Object without Score Distillation (ICLR 2025 · similarity 0.85)
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.
