Research map / 3D vision / 3D generation
Reconstruction and Simulation of Elastic Objects with Spring-Mass 3D Gaussians
Licheng Zhong, Hong-Xing Yu, Jiajun Wu, Yunzhu Li
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 |
|---|---|---|---|
| replace MPM-based PAC-NeRF dynamics with a spring-mass model on Gaussians | Spring-Gaus · Synthetic future prediction | Chamfer Distance (mean): 17.94 -> 5.16 | Improved |
| replace PAC-NeRF with Spring-Gaus for dynamic reconstruction | Spring-Gaus · Synthetic dynamic reconstruction | Chamfer Distance (mean): 2.11 -> 0.85 | Improved |
| add soft vector eta to modulate number of spring connections (ablation) | Spring-Gaus · Heterogeneous object, future prediction | Chamfer Distance: 13.28 -> 2.04 | Improved |
| add soft vector eta (ablation) | Spring-Gaus · Heterogeneous object, dynamic reconstruction | PSNR: 25.36 -> 27.08 | Improved |
| optimize per-anchor stiffness instead of a single global stiffness (ablation) | Spring-Gaus · Heterogeneous object, dynamic reconstruction | Chamfer Distance: 3.22 -> 0.18 | Improved |
| per-anchor stiffness instead of single k on future prediction (ablation) | Spring-Gaus · Heterogeneous object, future prediction | PSNR: 14.45 -> 17.63 | Improved |
| Spring-Gaus vs PAC-NeRF on Torus future prediction | Spring-Gaus · Synthetic future prediction, Torus | EMD / PSNR: EMD 0.087 vs 0.055; PSNR 16.83 vs 17.46 | Mixed |
Closest papers by meaning
- OmniPhysGS: 3D Constitutive Gaussians for General Physics-Based Dynamics Generation (ICLR 2025 · similarity 0.89)
- Physically Compatible 3D Object Modeling from a Single Image (NeurIPS 2024 · similarity 0.87)
- PhysGaussian: Physics-Integrated 3D Gaussians for Generative Dynamics (CVPR 2024 · similarity 0.86)
- GIC: Gaussian-Informed Continuum for Physical Property Identification and Simulation (NeurIPS 2024 · similarity 0.86)
- Motion Decoupled 3D Gaussian Splatting for Dynamic Object Representation (AAAI 2025 · similarity 0.85)
- Building Interactable Replicas of Complex Articulated Objects via Gaussian Splatting (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.
