Research map / 3D vision / Gaussian splatting
DynamicTree: Interactive Real Tree Animation via Sparse Voxel Spectrum
Yaokun Li, Lihe Ding, Xiao Chen, Guang Tan, Tianfan Xue
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 |
|---|---|---|---|
| use feed-forward sparse voxel spectrum generation instead of VDM-based 4D generation (SV4D 2.0) | SV4D 2.0 · Real-tree 3D animation (13 real trees) | CLIP-T distance: 0.0057 -> 0.0021 | Improved |
| use modal analysis on the generated spectrum instead of MPM simulation (PhysGaussian) | PhysGaussian · Interactive simulation | Simulation time (ms/frame): 1,800 -> 18.22 | Improved |
| use modal analysis on the generated spectrum instead of PhysFlow | PhysFlow · Interactive simulation | Simulation time (ms/frame): 15,600 -> 18.22 | Improved |
| DynamicTree versus PhysGaussian on structural consistency in user study | PhysGaussian · Interactive simulation (user study) | 3D structural consistency: 36.2% (PhysGaussian) vs 29.8% (ours) | Mixed |
| DynamicTree versus traditional hierarchical tree simulation (Weber) on clean meshes | Weber [58] · User study | Overall preference: 42.14% -> 57.86% | Improved |
| add Local Spectrum Smoothness loss with joint training from the start (ablation) | DynamicTree · real tree animation | visual artifacts (qualitative) | Mixed |
| two-stage training: diffusion loss first, then add LSS loss (ablation) | DynamicTree · real tree animation | visual artifacts and generalization (qualitative) | Improved |
| increase sparse voxel spectrum resolution from 32^3 to 128^3 (ablation) | DynamicTree · Voxel spectrum resolution ablation | CLIP-I distance: 0.0097 -> 0.0039 | Improved |
| increase resolution from 128^3 to 256^3 (ablation) | DynamicTree · Voxel spectrum resolution ablation | CLIP-I distance: 0.0039 -> 0.0037 (training time 85h -> 156h) | No clear effect |
| increase resolution from 128^3 to 512^3 (ablation) | DynamicTree · Voxel spectrum resolution ablation | CLIP-I distance: 0.0039 -> 0.0056 | Hurt |
Closest papers by meaning
- SVDTree: Semantic Voxel Diffusion for Single Image Tree Reconstruction (CVPR 2024 · similarity 0.85)
- TreeMeshGPT: Artistic Mesh Generation with Autoregressive Tree Sequencing (CVPR 2025 · similarity 0.85)
- Real-time Photorealistic Dynamic Scene Representation and Rendering with 4D Gaussian Splatting (ICLR 2024 · similarity 0.85)
- Dynamic Gaussians Mesh: Consistent Mesh Reconstruction from Dynamic Scenes (ICLR 2025 · similarity 0.85)
- 4D Gaussian Splatting for Real-Time Dynamic Scene Rendering (CVPR 2024 · similarity 0.85)
- Align Your Gaussians: Text-to-4D with Dynamic 3D Gaussians and Composed Diffusion Models (CVPR 2024 · 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.
