Research map / 3D vision / 3D generation
V2M4: 4D Mesh Animation Reconstruction from a Single Monocular Video
Jianqi Chen, Biao Zhang, Xiangjun Tang, Peter Wonka
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 PSO camera search (ablation) | V2M4 · V2M4 benchmark (Simple) | FVD: 1090.86 -> 825.59 | Improved |
| add gradient descent camera refinement (ablation) | V2M4 · V2M4 benchmark (Simple) | FVD: 1086.14 -> 825.59 | Improved |
| add negative condition embedding optimization (mesh appearance refinement) (ablation) | V2M4 · V2M4 benchmark (Simple) | FVD: 1023.60 -> 825.59 | Improved |
| add DUSt3R camera pose prior (ablation) | V2M4 · V2M4 benchmark (Simple) | FVD: 756.29 -> 825.59 | Mixed |
| replace DUSt3R with VGGT (ablation) | V2M4 · V2M4 benchmark (Simple) | DreamSim: 0.0688 -> 0.0631 | Mixed |
| use Hunyuan3D-2.0 instead of TRELLIS as base 3D generator (ablation) | V2M4 · V2M4 benchmark (All) | FVD: 576.73 -> 377.14 | Improved |
| apply V2M4 pipeline over naive per-frame TRELLIS | TRELLIS · V2M4 benchmark (All) | FVD: 1014.45 -> 576.73 | Improved |
| increase ARAP loss weight from 1 to 10 (ablation) | V2M4 · V2M4 benchmark (Simple) | FVD: 825.59 -> 928.54 | Hurt |
| inject 10x Gaussian noise during mesh generation (ablation) | V2M4 · V2M4 benchmark (Simple) | FVD: 825.59 -> 2410.03 | Hurt |
| inject 0.1x noise during mesh generation (ablation) | V2M4 · V2M4 benchmark (Simple) | CLIP: 0.9259 -> 0.9289 | No clear effect |
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.
- DreamMesh4D — DreamMesh4D: Video-to-4D Generation with Sparse-Controlled Gaussian-Mesh Hybrid Representation (NeurIPS 2024) · better on 7 of 8 shared results
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
- DreamMesh4D: Video-to-4D Generation with Sparse-Controlled Gaussian-Mesh Hybrid Representation (NeurIPS 2024 · similarity 0.91)
- Vidu4D: Single Generated Video to High-Fidelity 4D Reconstruction with Dynamic Gaussian Surfels (NeurIPS 2024 · similarity 0.88)
- EG4D: Explicit Generation of 4D Object without Score Distillation (ICLR 2025 · similarity 0.88)
- Animate3D: Animating Any 3D Model with Multi-view Video Diffusion (NeurIPS 2024 · similarity 0.87)
- Fast 4D Mesh Generation by Spatio-Temporal Attention Chains (NeurIPS 2026 · similarity 0.87)
- Consistent4D: Consistent 360° Dynamic Object Generation from Monocular Video (ICLR 2024 · 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.
