Research map / 3D vision / Human pose estimation
KASALv2: Fully Automatic 3D Rotational Symmetry Classification and Axis Localization
Mengxin Zhang, Yulin Wang, Chen Luo, Yongzhe Li, Yijun 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 |
|---|---|---|---|
| train FoundationPose with KASALv2 symmetry priors | FoundationPose · IC-BIN | BOP accuracy: 63.8 -> 64.7 | Improved |
| train FoundationPose with KASALv2 symmetry priors | FoundationPose · YCB-V | BOP accuracy: 87.6 -> 87.7 | No clear effect |
| train FoundationPose with KASALv2 symmetry priors | FoundationPose · ITODD | BOP accuracy: 65.0 -> 65.8 | Improved |
| train FoundationPose with KASALv2 symmetry priors | FoundationPose · LM-O | BOP accuracy: 71.6 -> 71.9 | Improved |
| use stricter BOP re-annotation instead of original labels | KASALv2 evaluation · BOP (50 symmetric objects) | acc_T: 80.00% -> 84.00% | Improved |
| KASALv2 localization vs Wang et al. KASAL | KASALv2 · BOP | e_ADI/d: 0.00290 -> 0.00256 | Improved |
| KASALv2 localization vs Wang et al. KASAL | KASALv2 · DSRSTO | e_ADI/d: 0.00172 -> 0.00212 | Hurt |
| vary spherical-cap radius away from 10 degrees (ablation) | KASALv2 · DSRSTO | acc_T: 81.48% at r=10 deg | Hurt |
| use smaller cap radius (<=8 deg) (ablation) | KASALv2 · DSRSTO | acc_T | No clear effect |
| reduce candidate order set to {3,4} or {3,4,5} (ablation) | KASALv2 · DSRSTO | acc_T | Hurt |
| expand candidate order set to {3,4,5,6,7} or {3,4,5,6,7,9} (ablation) | KASALv2 · DSRSTO | runtime | Hurt |
| include the 2-fold term in the candidate order set (ablation) | KASALv2 · DSRSTO | acc_T | Hurt |
Closest papers by meaning
- Towards Symmetry-sensitive Pose Estimation: A Rotation Representation for Symmetric Object Classes (IJCV 2026 · similarity 0.90)
- 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)
- Symmetry-Robust 3D Orientation Estimation (ICML 2025 · similarity 0.87)
- Relaxed Rotational Equivariance via G-Biases in Vision (AAAI 2025 · similarity 0.85)
- R2Det: Exploring Relaxed Rotation Equivariance in 2D Object Detection (ICLR 2025 · similarity 0.84)
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.
