Research map / Point clouds / Shape and surface completion
PartDiffuser: Part-wise 3D Mesh Generation via Discrete Diffusion
Yichen Yang, Hong Li, Haodong Zhu, Linin Yang, Guojun Lei, Sheng Xu, Baochang Zhang
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 part-specific conditioning to global-only conditioning (ablation) | PartDiffuser · Objaverse | CD x10^3: 29.125 -> 17.813 | Improved |
| add global conditioning to parts-only conditioning (ablation) | PartDiffuser · Objaverse | CD x10^3: 54.728 -> 17.813 | Improved |
| add part-specific conditioning to global-only conditioning (ablation) | PartDiffuser · Objaverse | F1: 0.281 -> 0.343 | Improved |
| add global conditioning to parts-only conditioning (ablation) | PartDiffuser · Objaverse | F1: 0.239 -> 0.343 | Improved |
| increase sampling acceleration factor from k=1 to k=2 (steps 1024 to 512) (ablation) | PartDiffuser · Objaverse | CD x10^3: 17.813 -> 27.657 | Hurt |
| increase sampling acceleration factor from k=1 to k=4 (steps 1024 to 256) (ablation) | PartDiffuser · Objaverse | CD x10^3: 17.813 -> 44.720 | Hurt |
| increase sampling acceleration factor from k=1 to k=4 (ablation) | PartDiffuser · Objaverse | inference time (s): 63.8 -> 17.1 | Improved |
| part-wise semi-autoregressive diffusion vs MeshAnythingV2 | PartDiffuser · Objaverse | CD x10^3: 24.402 -> 17.813 | Improved |
| part-wise diffusion vs TreeMeshGPT | PartDiffuser · HSSD | CD x10^3: 4.522 -> 5.364 | Mixed |
| part-wise diffusion vs TreeMeshGPT | PartDiffuser · 3DFront | F1: 0.462 -> 0.453 | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- TreeMeshGPT — TreeMeshGPT: Artistic Mesh Generation with Autoregressive Tree Sequencing (CVPR 2025) · better on 6 of 8 shared results
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
- PartCrafter: Structured 3D Mesh Generation via Compositional Latent Diffusion Transformers (NeurIPS 2025 · similarity 0.91)
- PartGen: Part-level 3D Generation and Reconstruction with Multi-view Diffusion Models (CVPR 2025 · similarity 0.90)
- ARMesh: Autoregressive Mesh Generation via Next-Level-of-Detail Prediction (NeurIPS 2025 · similarity 0.90)
- AutoPartGen: Autoregressive 3D Part Generation and Discovery (NeurIPS 2025 · similarity 0.89)
- MeshAnything: Artist-Created Mesh Generation with Autoregressive Transformers (ICLR 2025 · similarity 0.89)
- Diffusion 3D Features (Diff3F): Decorating Untextured Shapes with Distilled Semantic Features (CVPR 2024 · 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.
