Research map / 3D vision / Human pose estimation
Algebraically Rigorous Quaternion Framework for the Neural Network Pose Estimation Problem
Chen Lin, Andrew J. Hanson, Sonya M. Hanson
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 to quaternion adjugate matrix instead of a lone quaternion (ablation) | PointNet pose regression · simulated random point clouds (N=100, sigma=0.1) | mean distance to target quaternion | Improved |
| train to quaternion adjugate instead of 6D rotation representation | PointNet pose regression · simulated random point clouds | mean distance to target quaternion | Improved |
| train to quaternion adjugate instead of A matrix | PointNet pose regression · simulated random point clouds | mean distance to target quaternion | No clear effect |
| train to quaternion adjugate instead of A matrix on multiple airplane models | PointNet pose regression · ModelNet40, 33 airplanes | angular error | Improved |
| train to quaternion adjugate instead of A matrix on a single airplane | PointNet pose regression · ModelNet40, single airplane | angular error: both around 1 degree | No clear effect |
| train to quaternion adjugate on single-source many-rotation setting | PointNet pose regression · single-source simulated clouds | mean angular error (degrees): 0.66 (6D) -> 0.44 (adjugate); A matrix 0.37 | Mixed |
| use optimal Horn quaternion rather than the simulation quaternion q_init as target (ablation) | adjugate training target · simulated point clouds sigma=0.1 | training reliability | Improved |
Closest papers by meaning
- Quaternion Equivariant Capsule Networks for 3D Point Clouds (ECCV 2020 · similarity 0.89)
- Dual Quaternion SE(3) Synchronization with Recovery Guarantees (ICML 2026 · similarity 0.89)
- 3D-Rotation-Equivariant Quaternion Neural Networks (ECCV 2020 · similarity 0.88)
- Asymptotically Stable Quaternion-valued Hopfield-structured Neural Network with Periodic Projection-based Supervised Learning Rules (NeurIPS 2025 · similarity 0.88)
- Modeling Human Motion with Quaternion-Based Neural Networks (IJCV 2019 · similarity 0.88)
- 3D Equivariant Pose Regression via Direct Wigner-D Harmonics Prediction (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.
