Research map / 3D vision / Human pose estimation
SCAPO: Self-Supervised Category-Level Articulated Pose Estimation from a Single 3D Observation
Can Zhang, Gim Hee Lee
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 shape variance regularization (ablation) | SCAPO · Synthetic HOI4D+Shape2Motion (mean) | segmentation / joint direction error / pivot error: 84.98 vs 80.15; 3.89 vs 10.63; 0.075 vs 0.175 | Improved |
| add joint direction alignment loss (ablation) | SCAPO · Synthetic HOI4D+Shape2Motion (mean) | joint direction error: 8.54 -> 3.89 | Improved |
| add joint direction alignment loss (ablation) | SCAPO · Synthetic HOI4D+Shape2Motion (mean) | joint pivot error: 0.086 -> 0.075 | Improved |
| add joint direction alignment loss (ablation) | SCAPO · Synthetic HOI4D+Shape2Motion (mean) | segmentation: 84.21 -> 84.98 | No clear effect |
| add joint-to-shape proximity loss (ablation) | SCAPO · Synthetic HOI4D+Shape2Motion (mean) | joint pivot error: 0.137 -> 0.075 | Improved |
| add joint-to-shape proximity loss (ablation) | SCAPO · Synthetic HOI4D+Shape2Motion (mean) | segmentation: 82.77 -> 84.98 | Improved |
| add joint-to-shape proximity loss (ablation) | SCAPO · Synthetic HOI4D+Shape2Motion (mean) | joint direction error: 4.73 -> 3.89 | Improved |
| add joint-to-boundary attraction loss (ablation) | SCAPO · Synthetic HOI4D+Shape2Motion (mean) | joint pivot error: 0.113 -> 0.075 | Improved |
| add joint-to-boundary attraction loss (ablation) | SCAPO · Synthetic HOI4D+Shape2Motion (mean) | segmentation / joint direction error: 84.50 -> 84.98; 4.12 -> 3.89 | No clear effect |
| self-supervised SCAPO instead of OP-Align | SCAPO · Synthetic HOI4D+Shape2Motion (mean) | mean segmentation IoU / joint direction / pivot / rotation: 82.24 -> 84.98; 4.60 -> 3.89; 0.104 -> 0.075; 6.73 -> 5.78 | Improved |
| self-supervised SCAPO instead of OP-Align | SCAPO · Synthetic HOI4D+Shape2Motion (mean) | mean translation error: 0.115 -> 0.086 | Improved |
| self-supervised SCAPO instead of EAP | SCAPO · Synthetic HOI4D+Shape2Motion (mean) | mean translation error: 0.067 -> 0.086 | Hurt |
Closest papers by meaning
- SchemaPose: RGB-Based Category-Level Pose Estimation with Parametric Category Schema (NeurIPS 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 0.88)
- Unsupervised Learning of Category-Level 3D Pose from Object-Centric Videos (CVPR 2024 · similarity 0.88)
- Emergence of a Shared Canonical Object Frame from In-the-Wild Videos (NeurIPS 2026 · similarity 0.87)
- S3O: A Dual-Phase Approach for Reconstructing Dynamic Shape and Skeleton of Articulated Objects from Single Monocular Video (ICML 2024 · similarity 0.87)
- SINGAPO: Single Image Controlled Generation of Articulated Parts in Objects (ICLR 2025 · similarity 0.87)
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.
