Research map / 3D vision / 3D generation
Vivid4D: Improving 4D Reconstruction from Monocular Video by Video Inpainting
Jiaxin Huang, Sheng Miao, Bangbang Yang, Yuewen Ma, Yiyi Liao
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 anchor video condition (ablation) | video inpainting model · OpenVid-1M 5K videos (inpainting) | PSNR: 25.34 -> 27.22 | Improved |
| add anchor video condition (ablation) | video inpainting model · OpenVid-1M 5K videos (inpainting) | FVD: 18.99 -> 14.30 | Improved |
| add direct depth supervision loss (ablation) | no-augmentation baseline · iPhone + HyperNeRF (ablation) | mPSNR: 16.04 -> 16.12 | Improved |
| add view warping to synthesize augmented training views (ablation) | no-augmentation baseline · iPhone + HyperNeRF (ablation) | mPSNR: 16.04 -> 16.44 | Improved |
| add video inpainting after view warping (ablation) | warping only · iPhone + HyperNeRF (ablation) | mPSNR: 16.44 -> 16.80 | Improved |
| use iterative augmentation (N=6, j=6) instead of single-shot warping (ablation) | view augmentation · iPhone + HyperNeRF (ablation) | mPSNR: 15.80 -> 16.80 | Improved |
| increase iterations from j=2 to j=6 (ablation) | view augmentation · iPhone + HyperNeRF (ablation) | mPSNR: 16.41 -> 16.80 | Improved |
| single-shot augmentation versus no view augmentation (ablation) | reconstruction · iPhone + HyperNeRF (ablation) | mPSNR: 16.04 -> 15.80 | Hurt |
| use IV robust loss instead of L1 (ablation) | reconstruction loss · iPhone + HyperNeRF (ablation) | mPSNR: 16.50 -> 16.80 | Improved |
| use motion field backbone instead of 4D GS (ablation) | reconstruction backbone · iPhone + HyperNeRF (ablation) | mPSNR: 16.16 -> 16.80 | Improved |
| use Depth Anything V2 instead of DepthCrafter (ablation) | depth estimator · iPhone dataset (depth model ablation) | mPSNR: 16.76 -> 16.80 | No clear effect |
| replace Shape of Motion with the proposed method | Shape of Motion · iPhone dataset (All) | mLPIPS: 0.5292 -> 0.4930 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ViewCrafter — ViewCrafter: Taming Video Diffusion Models for High-fidelity Novel View Synthesis (TPAMI 2025) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Reconstruct, Inpaint, Test-Time Finetune: Dynamic Novel-view Synthesis from Monocular Videos (NeurIPS 2025 · similarity 0.92)
- CAT4D: Create Anything in 4D with Multi-View Video Diffusion Models (CVPR 2025 · similarity 0.90)
- DreamScene4D: Dynamic Multi-Object Scene Generation from Monocular Videos (NeurIPS 2024 · similarity 0.88)
- GFlow: Recovering 4D World from Monocular Video (AAAI 2025 · similarity 0.88)
- One View Is Enough: In-the-Wild Monocular Pretraining for Novel View Generation (NeurIPS 2026 · similarity 0.87)
- ViDAR: Video Diffusion-Aware 4D Reconstruction From Monocular Inputs (NeurIPS 2025 · 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.
