Research map / 3D vision / Human pose estimation
Category-Level Articulated Object Pose Estimation via Pose–Shape Hypothesis Generation and Verification
Kaifeng Tang, Chi Xu, Xin Ao, Yuting Ge, Tingrui Guo, Jun Zhou
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 |
|---|---|---|---|
| model joint pose-shape distribution and verify with instance-specific joints (vs pose-only generation) (ablation) | pose-shape hypothesis generator · HOI4D | mAP 5deg5cm (%): 51.1 -> 59.2 | Improved |
| add shape distribution without verification (ablation) | pose-only generator · HOI4D | mAP 5deg5cm (%): 51.1 -> 48.3 | Hurt |
| add joint verification to pose-only hypotheses (ablation) | pose-only generator · HOI4D | mAP 5deg5cm (%): 51.1 -> 53.8 | Improved |
| add SDF shape-consistency verification and denoising (ablation) | pose-shape generator + joint verification · HOI4D | mAP 5deg5cm (%): 59.2 -> 61.0 | Improved |
| add shape verification without denoising (SDF) (ablation) | pose-shape generator + joint verification · HOI4D | mAP 5deg5cm (%): 59.2 -> 54.4 | Hurt |
| add part-level point cloud denoising to SDF shape verification (ablation) | verification module · HOI4D | mAP 5deg5cm (%): 54.4 -> 61.0 | Improved |
| use SDF scoring instead of unidirectional Chamfer Distance for shape verification (ablation) | verification module · HOI4D | mAP 5deg5cm (%): 60.2 -> 61.0 | Improved |
| use generation-and-verification instead of regression or tracking baselines | full method · HOI4D | mAP 5deg5cm (%): 44.7 (PPF-Tracker) -> 61.0 | Improved |
| use single-view method versus sequence-based PPF-Tracker on synthetic data | full method · PM-Videos | Terr (m): Rerr 4.1 -> 3.0; Terr 0.049 -> 0.059 (Laptop) | Mixed |
| use learned SDF shape decoder vs CARTO decoder | full method · HOI4D shape reconstruction | CD-L2: 0.054 -> 0.016 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CARTO — CARTO: Category and Joint Agnostic Reconstruction of ARTiculated Objects (CVPR 2023) · 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
- ComPose: A Unified Completion-Pose Framework for Robust Category-Level Object Pose Estimation (CVPR 2026 · similarity 0.90)
- PriorPose: Reference-Guided Joint Deformation and Alignment for Category-Level Object Pose Estimation (ECCV 2026 · similarity 0.90)
- Self-Supervised Category-Level Articulated Object Pose Estimation with Part-Level SE(3) Equivariance (ICLR 2023 · similarity 0.90)
- Rethinking Correspondence-based Category-Level Object Pose Estimation (CVPR 2025 · similarity 0.89)
- 3D-Aware Hypothesis & Verification for Generalizable Relative Object Pose Estimation (ICLR 2024 · similarity 0.89)
- EfficientCAPER: An End-to-End Framework for Fast and Robust Category-Level Articulated Object Pose Estimation (NeurIPS 2024 · 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.
