Research map / 3D vision / 3D generation
ArtLLM: Generating Articulated Assets via 3D LLM
Penghao Wang, Siyuan Xie, Hongyu Yan, Xianghui Yang, Jingwei Huang, Chunchao Guo, Jiayuan Gu
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 quantized discrete tokens for layout and joints (ablation) | ArtLLM · PartNet-Mobility ablation (144 objects, PartNet-Mobility training only) | part IoU: 0.352 -> 0.473 | Improved |
| add quantized discrete tokens (ablation) | ArtLLM · PartNet-Mobility ablation (144 objects, PartNet-Mobility training only) | joint axis error (lower better): 0.277 -> 0.141 | Improved |
| add multi-task SFT (ablation) | ArtLLM · PartNet-Mobility ablation (144 objects, PartNet-Mobility training only) | joint type accuracy: 0.825 -> 0.898 | Improved |
| add multi-task SFT (ablation) | ArtLLM · PartNet-Mobility ablation (144 objects, PartNet-Mobility training only) | joint axis error (lower better): 0.289 -> 0.141 | Mixed |
| add multi-task SFT (ablation) | ArtLLM · PartNet-Mobility ablation (144 objects, PartNet-Mobility training only) | joint range IoU: 0.510 -> 0.582 | Improved |
| add random scaling and rotation augmentation (ablation) | ArtLLM · PartNet-Mobility ablation (144 objects, PartNet-Mobility training only) | part IoU: 0.412 -> 0.473 | Improved |
| add augmentation (ablation) | ArtLLM · PartNet-Mobility ablation (144 objects, PartNet-Mobility training only) | graph accuracy: 0.754 -> 0.780 | Improved |
| add multi-stage training with part-layout pretraining (ablation) | ArtLLM · PartNet-Mobility ablation (144 objects, PartNet-Mobility training only) | joint pivot error (lower better): 0.175 -> 0.135 | Improved |
| add multi-stage training (ablation) | ArtLLM · PartNet-Mobility ablation (144 objects, PartNet-Mobility training only) | joint range IoU: 0.511 -> 0.582 | Improved |
| ArtLLM vs retrieval-based Articulate-Anything | ArtLLM · PartNet-Mobility (7 categories, 77 objects) | part mIoU: 0.3381 -> 0.6884 | Improved |
| ArtLLM vs Articulate-Anything | ArtLLM · PartNet-Mobility (7 categories, 77 objects) | time (s): 522 -> 19 | Improved |
| retrain SINGAPO on our larger dataset (ablation) | SINGAPO · PartNet-Mobility (7 categories, 77 objects) | joint type accuracy: 0.7649 -> 0.9065 | Improved |
Closest papers by meaning
- Articulate-Anything: Automatic Modeling of Articulated Objects via a Vision-Language Foundation Model (ICLR 2025 · similarity 0.91)
- Rigel3D: Rig-aware Latents for Animation-Ready 3D Asset Generation (NeurIPS 2026 · similarity 0.91)
- MeshArt: Generating Articulated Meshes with Structure-Guided Transformers (CVPR 2025 · similarity 0.90)
- Building Interactable Replicas of Complex Articulated Objects via Gaussian Splatting (ICLR 2025 · similarity 0.89)
- MagicArticulate: Make Your 3D Models Articulation-Ready (CVPR 2025 · similarity 0.89)
- URDF-Anything: Constructing Articulated Objects with 3D Multimodal Language Model (NeurIPS 2025 · similarity 0.89)
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.
