Research map / Point clouds / Shape and surface completion
Rethinking 3D Shape Generation: Diffusion over Superquadrics
Zhiyang Liu, Li Wanze, Yuwei Wu, Chengran Yuan, JIAWEI SUN, Rui Zheng, Marcelo H Ang Jr
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 existence token (ablation) | DoSs · ShapeNet Chair | 1-NNA@CD: 93.32 -> 58.36 | Improved |
| add existence token (ablation) | DoSs · ShapeNet Chair | COV@CD: 19.91 -> 48.41 | Improved |
| add 6D rotation representation instead of Euler angles (ablation) | DoSs · ShapeNet Chair | 1-NNA@CD: 58.36 -> 57.57 | Improved |
| add 6D rotation representation instead of Euler angles (ablation) | DoSs · ShapeNet Chair | COV@CD: 48.41 -> 50.37 | Improved |
| sort primitives by y-position instead of random order (ablation) | DoSs · ShapeNet Chair | 1-NNA@CD: 57.57 -> 55.28 | Improved |
| sort primitives by y-position instead of random order (ablation) | DoSs · ShapeNet Chair | COV@CD: 50.37 -> 50.93 | Improved |
| sort primitives by volume instead of random order (ablation) | DoSs · ShapeNet Chair | 1-NNA@CD: 57.57 -> 53.80 | Improved |
| sort primitives by volume instead of random order (ablation) | DoSs · ShapeNet Chair | COV@CD: 50.37 -> 51.31 | Improved |
| train one multi-class model instead of category-specific models | DoSs · ShapeNet Chair | 1-NNA@CD: 53.80 -> 55.70 | No clear effect |
| train one multi-class model instead of category-specific models | DoSs · ShapeNet Airplane | 1-NNA@CD: 61.92 -> 60.20 | Improved |
| train one multi-class model instead of category-specific models | DoSs · ShapeNet Car | 1-NNA@CD: 57.50 -> 61.11 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 3 on most shared results.
- DiT-3D — DiT-3D: Exploring Plain Diffusion Transformers for 3D Shape Generation (NeurIPS 2023) · better on 1 of 6 shared results
- LION — LION: Latent Point Diffusion Models for 3D Shape Generation (NeurIPS 2022) · better on 2 of 6 shared results
- TIGER — TIGER: Time-Varying Denoising Model for 3D Point Cloud Generation with Diffusion Process (CVPR 2024) · 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
- Scaling Diffusion Mamba with Bidirectional SSMs for Efficient 3D Shape Generation (AAAI 2025 · similarity 0.89)
- Single Mesh Diffusion Models with Field Latents for Texture Generation (CVPR 2024 · similarity 0.89)
- Neural Point Cloud Diffusion for Disentangled 3D Shape and Appearance Generation (CVPR 2024 · similarity 0.89)
- ShapeShifter: 3D Variations Using Multiscale and Sparse Point-Voxel Diffusion (CVPR 2025 · similarity 0.88)
- Sin3DM: Learning a Diffusion Model from a Single 3D Textured Shape (ICLR 2024 · similarity 0.88)
- GPLD3D: Latent Diffusion of 3D Shape Generative Models by Enforcing Geometric and Physical Priors (CVPR 2024 · similarity 0.88)
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.
