Research map / Detection and segmentation / Open-set and few-shot detection
GAPartNet: Cross-Category Domain-Generalizable Object Perception and Manipulation via Generalizable and Actionable Parts
Haoran Geng, Helin Xu, Chengyang Zhao, Chao Xu, Li Yi, Siyuan Huang, He Wang
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 |
|---|---|---|---|
| Use GAPartNet domain-generalizable segmentation | 3D part segmentation · GAPartNet seen and unseen categories | AP50: 76.5 vs 68.8 (seen); 37.2 vs 27.7 (unseen) | Improved |
| Use domain-invariant perception with GAPart poses | Part pose estimation · GAPartNet unseen categories | 3D mIoU: 36.3 -> 40.6 | Improved |
| Use GAPart-pose-based manipulation heuristics | Cross-category manipulation · SAPIEN unseen drawer, door, handle, and button tasks | success rate: drawer 90.0; door 55.0; handle 85.0; button 95.0 | Improved |
| add domain adversarial learning (ablation) | GAPart segmentation · Seen categories | AP50: 71.1 -> 70.6 | Hurt |
| add domain adversarial learning (ablation) | GAPart segmentation · Unseen categories | AP50: 28.1 -> 29.8 | Improved |
| add part-oriented feature query (ablation) | Domain-adversarial GAPart segmentation · Seen categories | AP50: 70.6 -> 71.6 | Improved |
| add part-oriented feature query (ablation) | Domain-adversarial GAPart segmentation · Unseen categories | AP50: 29.8 -> 32.3 | Improved |
| add multi-resolution features (ablation) | Part-oriented adversarial GAPart segmentation · Seen categories | AP50: 71.6 -> 73.7 | Improved |
| add multi-resolution features (ablation) | Part-oriented adversarial GAPart segmentation · Unseen categories | AP50: 32.3 -> 35.0 | Improved |
| add distribution-balancing focal loss (ablation) | Part-oriented multi-resolution adversarial GAPart segmentation · Seen categories | AP50: 73.7 -> 76.5 | Improved |
| add distribution-balancing focal loss (ablation) | Part-oriented multi-resolution adversarial GAPart segmentation · Unseen categories | AP50: 35.0 -> 37.2 | Improved |
Closest papers by meaning
- PartManip: Learning Cross-Category Generalizable Part Manipulation Policy From Point Cloud Observations (CVPR 2023 · similarity 0.92)
- Learning Part-Aware Dense 3D Feature Field For Generalizable Articulated Object Manipulation (ICLR 2026 · similarity 0.89)
- CAP-Net: A Unified Network for 6D Pose and Size Estimation of Categorical Articulated Parts from a Single RGB-D Image (CVPR 2025 · similarity 0.87)
- PartNeXt: A Next-Generation Dataset for Fine-Grained and Hierarchical 3D Part Understanding (NeurIPS 2025 · similarity 0.86)
- IAAO: Interactive Affordance Learning for Articulated Objects in 3D Environments (CVPR 2025 · similarity 0.86)
- General Articulated Objects Manipulation in Real Images via Part-Aware Diffusion Process (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.
