Research map / 3D vision / Human pose estimation
LEADER: Learning Reliable Local-to-Global Correspondences for LiDAR Relocalization
Jianshi Wu, Minghang Zhu, Dunqiang Liu, Wen Li, Sheng Ao, Siqi Shen, Chenglu Wen, Cheng Wang
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 |
|---|---|---|---|
| add Spatial Transformation (cylindrical projection) (ablation) | LEADER without TRR · NCLT | mean position error (m): 0.95 -> 0.59 | Improved |
| add TRR loss in training (ablation) | LEADER · NCLT | mean position error (m): 0.59 -> 0.35 | Improved |
| add TRR weighting at test time (ablation) | LEADER · NCLT | mean position error (m): 0.35 -> 0.31 | Improved |
| add Spatial Transformation under yaw-perturbed input (ablation) | LEADER · NCLT yaw-perturbed | mean position error (m): 6.59 -> 0.31 | Improved |
| add TRR-train without ST (ablation) | LEADER · NCLT | mean position error (m): 0.95 -> 0.89 | Improved |
| add TRR-train without ST on yaw-perturbed input (ablation) | LEADER · NCLT yaw-perturbed | mean position error (m): 8.04 -> 8.67 | Hurt |
| use cylindrical instead of spherical projection (ablation) | LEADER · NCLT | mean position error (m): 0.35 -> 0.31 | Improved |
| use TRR loss instead of mean Euclidean loss (ablation) | LEADER · NCLT | mean position error (m): 0.59 -> 0.31 | Improved |
| use TRR loss instead of Matching loss (ablation) | LEADER · NCLT | mean position error (m): 0.43 -> 0.31 | Improved |
| add TRR loss to SGLoc (ablation) | SGLoc · NCLT | mean orientation error (deg): 3.54 -> 2.61 | Improved |
| add TRR loss to SGLoc (ablation) | SGLoc · NCLT | mean position error (m): 1.83 -> 1.76 | Improved |
| use only the 0-25% most reliable points vs 75-100% least reliable (ablation) | LEADER · NCLT | mean position error (m): 2.39 -> 0.32 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- LightLoc — LightLoc: Learning Outdoor LiDAR Localization at Light Speed (CVPR 2025) · better on 6 of 6 shared results
- DiffLoc — DiffLoc: Diffusion Model for Outdoor LiDAR Localization (CVPR 2024) · better on 3 of 4 shared results
- HypLiLoc — HypLiLoc: Towards Effective LiDAR Pose Regression With Hyperbolic Fusion (CVPR 2023) · better on 1 of 1 shared result
- SGLoc — SGLoc: Scene Geometry Encoding for Outdoor LiDAR Localization (CVPR 2023) · better on 1 of 1 shared result
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)
- SOAR: Regression-based LiDAR Relocalization for UAVs (NeurIPS 2026 · similarity 0.88)
- LightLoc: Learning Outdoor LiDAR Localization at Light Speed (CVPR 2025 · similarity 0.87)
- GeoNLF: Geometry guided Pose-Free Neural LiDAR Fields (NeurIPS 2024 · similarity 0.86)
- Pre-training a Density-Aware Pose Transformer for Robust LiDAR-based 3D Human Pose Estimation (AAAI 2025 · similarity 0.86)
- GBlobs: Explicit Local Structure via Gaussian Blobs for Improved Cross-Domain LiDAR-based 3D Object Detection (CVPR 2025 · 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.
