Research map / 3D vision / 3D generation
UniPart: Part-Level 3D Generation with Unified 3D Geom-Seg Latents
Xufan He, Yushuang Wu, Xiaoyang Guo, Chongjie Ye, Jiaqing Zhou, Tianlei Hu, Xiaoguang Han, Dong Du
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 segmentation information to the VecSet latent space (Geom-Seg VecSet VAE) (ablation) | Hunyuan3D-2.1 VAE · VAE reconstruction | Chamfer Distance (x10^4): 1.29 -> 1.30 | No clear effect |
| generate part latents in a normalized canonical space (dual-space generation) (ablation) | UniPart part-level DiT · Part-level generation (qualitative) | part geometric fidelity | Improved |
| add local attention alongside global attention (ablation) | UniPart part-level DiT · Part-level generation (qualitative) | part cohesion and global consistency | Improved |
| add learnable space embedding to distinguish gcs and ncs tokens (ablation) | UniPart part-level DiT · Part-level generation (qualitative) | assembly success | Improved |
| UniPart vs X-Part on Chamfer Distance | part-level generation · Part-level generation geometry quality | Chamfer Distance (x10^2): 0.82 -> 0.72 | Improved |
| UniPart vs X-Part on F1@0.1 | part-level generation · Part-level generation geometry quality | F1@0.1 (x10^2): 88.90 -> 92.21 | Improved |
| UniPart vs X-Part on IoU | part-level generation · Part-level generation geometry quality | IoU (x10^2): 31.95 -> 21.99 | Hurt |
| joint geometry and segmentation generation on structurally complex inputs | UniPart whole-level DiT · Failure cases | output validity | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- XPart — X-Part: High Fidelity And Structure Coherent Shape Decomposition And Completion (CVPR 2026) · better on 2 of 3 shared results
- HoloPart — HoloPart: Generative 3D Part Amodal Segmentation (ICLR 2026) · better on 2 of 2 shared results
- PartCrafter — PartCrafter: Structured 3D Mesh Generation via Compositional Latent Diffusion Transformers (NeurIPS 2025) · better on 2 of 2 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)
- AutoPartGen: Autoregressive 3D Part Generation and Discovery (NeurIPS 2025 · similarity 0.88)
- Generative 3D Part Assembly via Part-Whole-Hierarchy Message Passing (CVPR 2024 · similarity 0.88)
- SegGraph: Leveraging Graphs of SAM Segments for Few-Shot 3D Part Segmentation (NeurIPS 2025 · similarity 0.88)
- PartNeXt: A Next-Generation Dataset for Fine-Grained and Hierarchical 3D Part Understanding (NeurIPS 2025 · 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.
