Research map / 3D vision / Face and shape
CompoNet: Learning to Generate the Unseen by Part Synthesis and Composition
Nadav Schor, Oren Katzir, Hao Zhang, Daniel Cohen-Or
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 |
|---|---|---|---|
| generate shapes by part synthesis and composition instead of whole-shape WGAN-GP | CompoNet · Chair, split (a) | % classified as unseen: 0.67 -> 0.76 | Improved |
| generate shapes by part synthesis and composition instead of whole-shape AE+GMM | CompoNet · Airplane, split (a) | % classified as unseen: 0.67 -> 0.8 | Improved |
| part-based generation instead of whole-shape VAE | CompoNet · Vase 2D, split (a) | % classified as unseen: 0.3 -> 0.46 | Improved |
| part-based generation instead of whole-shape WGAN-GP | CompoNet · Chair, split (b) | % classified as unseen: 0.75 -> 0.87 | Improved |
| part-based generation instead of whole-shape AE+GMM | CompoNet · Airplane, split (b) | % classified as unseen: 0.43 -> 0.86 | Improved |
| part-based generation instead of whole-shape AE+GMM | CompoNet · Chair, split (c) | % classified as unseen: 0.63 -> 0.55 | Hurt |
| part-based generation instead of whole-shape WGAN-GP | CompoNet · Chair, split (c) | % classified as unseen: 0.54 -> 0.55 | No clear effect |
| part-based generation instead of whole-shape WGAN-GP | CompoNet · Airplane, split (c) | % classified as unseen: 0.61 -> 0.65 | Improved |
| part-based generation instead of whole-shape AE+GMM | CompoNet · ShapeNet Chair | JSD to unseen set (lower better): 0.19 -> 0.02 | Improved |
| part-based generation instead of whole-shape AE+GMM | CompoNet · ShapeNet Airplane | JSD to unseen set (lower better): 0.14 -> 0.07 | Improved |
| part-based generation instead of whole-shape WGAN-GP | CompoNet · ShapeNet Chair | JSD to unseen set (lower better): 0.3 -> 0.02 | Improved |
| part-based generation instead of whole-shape WGAN-GP | CompoNet · ShapeNet Airplane | JSD to unseen set (lower better): 0.32 -> 0.07 | Improved |
Closest papers by meaning
- AutoPartGen: Autoregressive 3D Part Generation and Discovery (NeurIPS 2025 · similarity 0.88)
- PartCrafter: Structured 3D Mesh Generation via Compositional Latent Diffusion Transformers (NeurIPS 2025 · similarity 0.87)
- Compositional Scene Understanding through Inverse Generative Modeling (ICML 2025 · similarity 0.87)
- PartGen: Part-level 3D Generation and Reconstruction with Multi-view Diffusion Models (CVPR 2025 · similarity 0.87)
- Generative 3D Part Assembly via Part-Whole-Hierarchy Message Passing (CVPR 2024 · similarity 0.86)
- Make-A-Shape: a Ten-Million-scale 3D Shape Model (ICML 2024 · similarity 0.86)
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.
