Research map / 3D vision / Human pose estimation
HypLiLoc: Towards Effective LiDAR Pose Regression With Hyperbolic Fusion
Sijie Wang, Qiyu Kang, Rui She, Wei Wang, Kai Zhao, Yang Song, Wee Peng Tay
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 HypLiLoc instead of the strongest listed LiDAR pose-regression baselines | LiDAR pose regression · Oxford Radar Full-6 to Full-9 | mean translation and rotation error: Full-9 translation: 7.27 -> 3.45; rotation: 1.07 -> 0.84 | Improved |
| Use HypLiLoc instead of the strongest listed baselines | LiDAR pose regression · vReLoc Seq-05 to Seq-14 | median translation and rotation error: HypLiLoc achieves the best performance in 7 out of 8 metrics | Mixed |
| Add global graph attention to the base model (ablation) | HypLiLoc · Oxford Radar Full-8 | mean translation error (m): 9.78 -> 8.91 | Improved |
| Add a spherical-projection backbone after global graph attention (ablation) | HypLiLoc · Oxford Radar Full-8 | mean translation error (m): 8.91 -> 7.26 | Improved |
| Add the feature fusion block (ablation) | HypLiLoc · Oxford Radar Full-8 | mean translation error (m): 7.26 -> 6.19 | Improved |
| Add an unconstrained learnable metric (ablation) | HypLiLoc feature fusion · Oxford Radar Full-8 | mean translation error (m): 6.19 -> 5.82 | Improved |
| Add the hyperbolic branch to the Euclidean branch (ablation) | HypLiLoc · Oxford Radar Full-8 | mean translation error (m): 6.24 -> 5.82 | Improved |
| Add the Euclidean branch to the hyperbolic branch (ablation) | HypLiLoc · Oxford Radar Full-8 | mean translation error (m): 6.57 -> 5.82 | Improved |
| Use both 3D point-cloud and spherical-projection modalities (ablation) | HypLiLoc modality input · Oxford Radar Full-8 | mean translation error (m): 8.91 -> 5.82 | Improved |
| Replace spherical projection with BEV in the two-modality combination (ablation) | HypLiLoc modality input · Oxford Radar Full-8 | mean translation error (m): 5.82 -> 9.44 | Hurt |
| Add BEV as a third modality to 3D + spherical projection (ablation) | HypLiLoc modality input · Oxford Radar Full-8 | mean translation error (m): 5.82 -> 6.32 | Hurt |
| Constrain the learnable metric as a Riemannian metric rather than leave it unconstrained (ablation) | HypLiLoc feature fusion · Oxford Radar Full-8 | mean translation error (m): 5.82 -> 6.34 | Hurt |
Later papers that compared with this method
2 later accepted papers report HypLiLoc as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- LEADER: Learning Reliable Local-to-Global Correspondences for LiDAR Relocalization (CVPR 2026 · better on 1 of 1 shared result)
- DiffLoc: Diffusion Model for Outdoor LiDAR Localization (CVPR 2024 · better on 2 of 2 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- DiffLoc: Diffusion Model for Outdoor LiDAR Localization (CVPR 2024 · similarity 0.89)
- LEADER: Learning Reliable Local-to-Global Correspondences for LiDAR Relocalization (CVPR 2026 · similarity 0.88)
- LightLoc: Learning Outdoor LiDAR Localization at Light Speed (CVPR 2025 · similarity 0.88)
- Pre-training a Density-Aware Pose Transformer for Robust LiDAR-based 3D Human Pose Estimation (AAAI 2025 · similarity 0.85)
- GBlobs: Explicit Local Structure via Gaussian Blobs for Improved Cross-Domain LiDAR-based 3D Object Detection (CVPR 2025 · similarity 0.85)
- DiffLO: Semantic-Aware LiDAR Odometry with Diffusion-Based Refinement (CVPR 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.
