Research map / 3D vision / Human pose estimation
Implicit-PDF: Non-Parametric Representation of Probability Distributions on the Rotation Manifold
Kieran A Murphy, Carlos Esteves, Varun Jampani, Srikumar Ramalingam, Ameesh Makadia
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 |
|---|---|---|---|
| Use an implicit, non-parametric pose distribution instead of the compared parameterized distribution baselines | pose estimation · SYMSOL I | average log likelihood: -0.43 -> 4.1 | Improved |
| Use an implicit, non-parametric pose distribution instead of the compared parameterized distribution baselines | pose estimation · SYMSOL II | average log likelihood: 3.7 -> 7.57 | Improved |
| Return top-2 pose candidates rather than only the top-1 prediction | IPDF · ModelNet10-SO(3) | Acc@15°: 0.719 -> 0.868 | Improved |
| Return top-2 pose candidates rather than only the top-1 prediction | IPDF · ModelNet10-SO(3) | median angle error: 21.5 -> 4.9 | Improved |
| Use an implicit, non-parametric pose distribution instead of the compared methods | pose estimation · T-LESS | log likelihood: 8.8 -> 9.80 | Improved |
| Use an implicit, non-parametric pose distribution instead of the compared methods | pose estimation · T-LESS | spread (°): 3.44 -> 4.12 | Mixed |
| Use 50,000 rather than 10,000 training images per shape (ablation) | IPDF · SYMSOL I | average log likelihood | Improved |
| Add positional encoding to rotation queries (ablation) | IPDF · SYMSOL I discrete-symmetry shapes | average log likelihood | Improved |
| Add positional encoding to rotation queries (ablation) | IPDF · SYMSOL I cone and cylinder | average log likelihood | Mixed |
| Represent query rotations as matrices rather than axis-angle or quaternions (ablation) | IPDF · SYMSOL I | average log likelihood | Improved |
| Represent query rotations as Euler angles (ablation) | IPDF · SYMSOL I | average log likelihood | Hurt |
| Use randomly sampled rotations rather than an equivolumetric grid for training-time normalization (ablation) | IPDF · SYMSOL I | average log likelihood | No clear effect |
Closest papers by meaning
- Deep Directional Statistics: Pose Estimation with Uncertainty Quantification (ECCV 2018 · similarity 0.88)
- Towards Symmetry-sensitive Pose Estimation: A Rotation Representation for Symmetric Object Classes (IJCV 2026 · similarity 0.88)
- Image to Sphere: Learning Equivariant Features for Efficient Pose Prediction (ICLR 2023 · similarity 0.88)
- Pose-NDF: Modeling Human Pose Manifolds with Neural Distance Fields (ECCV 2022 · similarity 0.88)
- LEAP: Liberate Sparse-View 3D Modeling from Camera Poses (ICLR 2024 · similarity 0.86)
- Object Pose Estimation via the Aggregation of Diffusion Features (CVPR 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.
