Research map / 3D vision / 3D generation
PartField: Learning 3D Feature Fields for Part Segmentation and Beyond
Minghua Liu, Mikaela Angelina Uy, Donglai Xiang, Hao Su, Sanja Fidler, Nicholas Sharp, Jun Gao
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 PartNet 3D part labels as extra training triplets (ablation) | PartField (Objaverse 2D proposals) · PartObjaverse-Tiny | mIoU: 77.70 -> 77.90 | Improved |
| add hard negative sampling (ablation) | PartField (Objaverse 2D proposals) · PartObjaverse-Tiny | mIoU: 77.70 -> 78.90 | Improved |
| add hard negative sampling to Objaverse+PartNet (ablation) | PartField · PartObjaverse-Tiny | mIoU: 77.90 -> 79.20 | Improved |
| add PartNet triplets on top of hard negative sampling (ablation) | PartField · PartObjaverse-Tiny | mIoU: 78.90 -> 79.20 | Improved |
| use PartField instead of second-best SAMesh | SAMesh · PartObjaverse-Tiny | mIoU: 56.86 -> 79.18 | Improved |
| use PartField instead of SAMPart3D | SAMPart3D · PartNetE test (1906 shapes) | mIoU: 56.17 -> 59.10 | Improved |
| use feedforward model instead of per-shape optimization | SAMPart3D · PartObjaverse-Tiny | runtime: ~15min -> ~10s | Improved |
| use text-free 3D feature field instead of text-input methods | PartSLIP/Find3D · PartObjaverse-Tiny | mIoU: 31.54 -> 79.18 | Improved |
Later papers that compared with this method
7 later accepted papers report PartField as a baseline in their results tables; 5 of them report a better number on most of the results they share with it.
- Material Magic Wand: Material-Aware Grouping of 3D Parts in Untextured Meshes (CVPR 2026 · better on 3 of 3 shared results)
- PhysForge: Generating Physics-Grounded 3D Assets for Interactive Virtual World (ICML 2026 · better on 1 of 2 shared results)
- TEXTRIX: Latent Attribute Grid for Native Texture Generation and Beyond (CVPR 2026 · better on 1 of 2 shared results)
- GeoSAM2: Unleashing the Power of SAM2 for 3D Part Segmentation (CVPR 2026 · better on 2 of 2 shared results)
- XDen-1K: A Density Field Dataset of Real-World Objects (ECCV 2026 · better on 2 of 2 shared results)
- X-Part: High Fidelity And Structure Coherent Shape Decomposition And Completion (CVPR 2026 · better on 1 of 1 shared result)
- Articulated Object Reconstruction from Rest-State Observation (ECCV 2026 · 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
- PartSAM: A Scalable Promptable Part Segmentation Model Trained on Native 3D Data (ICLR 2026 · similarity 0.93)
- PartNet: A Large-Scale Benchmark for Fine-Grained and Hierarchical Part-Level 3D Object Understanding (CVPR 2019 · similarity 0.92)
- PartNeXt: A Next-Generation Dataset for Fine-Grained and Hierarchical 3D Part Understanding (NeurIPS 2025 · similarity 0.91)
- Learning Partonomic 3D Reconstruction from Image Collections (CVPR 2025 · similarity 0.90)
- PartDistill: 3D Shape Part Segmentation by Vision-Language Model Distillation (CVPR 2024 · similarity 0.89)
- PartCrafter: Structured 3D Mesh Generation via Compositional Latent Diffusion Transformers (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.
