Research map / 3D vision / Human pose estimation
Towards Symmetry-sensitive Pose Estimation: A Rotation Representation for Symmetric Object Classes
Andreas Kriegler, Csaba Beleznai, Margrit Gelautz
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 |
|---|---|---|---|
| replace trigonometric (canonic) with SARR representation | depth network · T-LESS | AR_C: 40.2 -> 45.2 | Improved |
| replace trigonometric (canonic) with SARR representation | depth network · ITODD | AR_C: 38.2 -> 41.3 | Improved |
| replace Euler angles with SARR | depth network · T-LESS | AR_C: 10.6 -> 45.2 | Improved |
| restrict rotation to canonic pose (canonic Euler) instead of ambiguous Euler labels | Euler representation · T-LESS | AR_C: 10.6 -> 27.7 | Improved |
| restrict rotation to canonic pose (canonic 6d) instead of ambiguous labels | 6d representation · T-LESS | AR_C: 4.62 -> 37.7 | Improved |
| add auxiliary symmetry classification task (dataset*) | SARR depth network · T-LESS | AR_C: 45.2 -> 44.7 | No clear effect |
| add auxiliary symmetry classification task (dataset*) | SARR depth network · ITODD | AR_C: 41.3 -> 42.9 | Improved |
| replace depth input with grayscale input (ablation) | SARR network (dataset scope) · ITODD SiSo | AR_C: 41.0 -> 2.8 | Hurt |
| use object scope instead of dataset scope (ablation) | SARR-RGB · T-LESS SiSo | AR_C: 47.3 -> 32.6 | Hurt |
| use symmetry scope instead of dataset scope (ablation) | SARR-RGB · T-LESS SiSo | AR_C: 47.3 -> 29.1 | Hurt |
| replace depth input with RGB input (ablation) | SARR network (object scope) · T-LESS SiSo | AR_C: 45.0 -> 32.6 | Hurt |
| use SARR depth network instead of SC6D | SC6D · ITODD SiSo | AR_B: 74.0 -> 62.1 | Hurt |
Closest papers by meaning
- KASALv2: Fully Automatic 3D Rotational Symmetry Classification and Axis Localization (CVPR 2026 · similarity 0.90)
- Learning to Detect 3D Symmetry From Single-View RGB-D Images With Weak Supervision (TPAMI 2022 · similarity 0.90)
- Distributional Estimation of 3D Object Orientation (NeurIPS 2026 · similarity 0.89)
- RECON: Robust symmetry discovery via Explicit Canonical Orientation Normalization (ICLR 2026 · similarity 0.89)
- Leveraging 3D Geometric Priors in 2D Rotation Symmetry Detection (CVPR 2025 · similarity 0.89)
- Orient Anything V2: Unifying Orientation and Rotation Understanding (NeurIPS 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.
