Research map / 3D vision / 3D generation
SK-Adapter: Skeleton-Based Structural Control for Native 3D Generation
Anbang Wang, AO Yuzhuo, Elliott (Shangzhe) Wu, Chi-Keung Tang
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 dedicated skeletal cross-attention instead of concatenating skeleton with text features (ablation) | SK-Adapter · TMS-eval, 8k subset | ReRigging Score: 0.5049 -> 0.2355 | Improved |
| add dedicated skeletal cross-attention (ablation) | SK-Adapter · TMS-eval, 8k subset | KD-DINO: 1.0914 -> 0.8616 | Improved |
| add topology-aware GRPE encoder (ablation) | SK-Adapter · TMS-eval, 8k subset | ReRigging Score: 0.2527 -> 0.2355 | Improved |
| add topology-aware GRPE encoder (ablation) | SK-Adapter · TMS-eval, 8k subset | CLIP Score: 26.14 -> 26.11 | No clear effect |
| add topology-aware GRPE encoder (ablation) | SK-Adapter · TMS-eval, 8k subset | KD-DINO: 0.7865 -> 0.8616 | Hurt |
| add topology-aware GRPE encoder (ablation) | SK-Adapter · TMS-eval, 8k subset | PickScore: 20.90 -> 20.94 | No clear effect |
| native 3D skeletal token injection vs 2D-lifting SKDream | SK-Adapter · TMS-eval | ReRigging Score: 0.2818 -> 0.2228 | Improved |
| trained adapter with frozen backbone vs training-free SpaceControl | SK-Adapter · TMS-eval | KD-DINO: 1.7821 -> 0.7778 | Improved |
| SK-Adapter vs baselines on 'other objects' category | SK-Adapter · TMS-eval (other objects) | PickScore: 20.45 (SpaceControl) vs 20.44 (SK-Adapter) | No clear effect |
Closest papers by meaning
- SKDream: Controllable Multi-view and 3D Generation with Arbitrary Skeletons (CVPR 2025 · similarity 0.89)
- SILSA: Sliding-Window Slice Latents for Topology-Preserving High-Resolution 3D Generation (NeurIPS 2026 · similarity 0.88)
- CraftsMan3D: High-fidelity Mesh Generation with 3D Native Diffusion and Interactive Geometry Refiner (CVPR 2025 · similarity 0.86)
- Structured 3D Latents for Scalable and Versatile 3D Generation (CVPR 2025 · similarity 0.86)
- Rigel3D: Rig-aware Latents for Animation-Ready 3D Asset Generation (NeurIPS 2026 · similarity 0.86)
- MIDGArD: Modular Interpretable Diffusion over Graphs for Articulated Designs (NeurIPS 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.
