Research map / 3D vision / Face and shape
Self-supervised Learning of Hybrid Part-aware 3D Representations of 2D Gaussians and Superquadrics
Zhirui Gao, Renjiao Yi, Yuhang Huang, Wei Chen, Chenyang Zhu, Kai Xu
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 opacity loss (ablation) | Ours (Block-level) · ShapeNet chair ablation | Chamfer distance: 5.72 -> 2.32 | Improved |
| add coverage loss (ablation) | Ours (Block-level) · ShapeNet chair ablation | Chamfer distance: 3.21 -> 2.32 | Improved |
| add adaptive strategy (ablation) | Ours (Block-level) · ShapeNet chair ablation | Chamfer distance: 3.16 -> 2.32 | Improved |
| add overlap loss (ablation) | Ours (Block-level) · ShapeNet chair ablation | Chamfer distance: 2.11 -> 2.32 | Hurt |
| add parsimony loss (ablation) | Ours (Block-level) · ShapeNet chair ablation | Chamfer distance: 1.91 -> 2.32 | Hurt |
| add parsimony loss (ablation) | Ours (Block-level) · ShapeNet chair ablation | number of primitives: 10.1 -> 6.6 | Improved |
| add overlap loss (ablation) | Ours (Point-level) · ShapeNet chair ablation | Chamfer distance: 1.34 -> 0.94 | Improved |
| increase parsimony weight from 0.01 to 0.1 (ablation) | Ours (Block-level) · DTU | Chamfer distance: 4.19 -> 6.46 | Hurt |
| increase initial primitives M from 4 to 16 (ablation) | Ours (Block-level) · DTU | Chamfer distance: 4.97 -> 3.99 | Improved |
| increase initial primitives M from 8 to 16 (ablation) | Ours (Point-level) · DTU | Chamfer distance: 1.05 -> 1.07 | No clear effect |
| use point-level instead of block-level representation | Ours · DTU | Chamfer distance: 4.19 -> 0.98 | Improved |
| use hybrid Gaussians+superquadrics instead of DBW | DBW · DTU | Chamfer distance: 4.76 -> 0.98 | Improved |
Closest papers by meaning
- Learning Partonomic 3D Reconstruction from Image Collections (CVPR 2025 · similarity 0.89)
- GaussianBlock: Building Part-Aware Compositional and Editable 3D Scene by Primitives and Gaussians (ICLR 2025 · similarity 0.88)
- GaussianUDF: Inferring Unsigned Distance Functions through 3D Gaussian Splatting (CVPR 2025 · similarity 0.88)
- ArticulatedGS: Self-supervised Digital Twin Modeling of Articulated Objects using 3D Gaussian Splatting (CVPR 2025 · similarity 0.88)
- PartCrafter: Structured 3D Mesh Generation via Compositional Latent Diffusion Transformers (NeurIPS 2025 · similarity 0.88)
- Evolving High-Quality Rendering and Reconstruction in a Unified Framework with Contribution-Adaptive Regularization (CVPR 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.
