Research map / 3D vision / Depth and stereo
OrienterNet: Visual Localization in 2D Public Maps With Neural Matching
Paul-Edouard Sarlin, Daniel DeTone, Tsun-Yi Yang, Armen Avetisyan, Julian Straub, Tomasz Malisiewicz, Samuel Rota Bulò, Richard Newcombe, Peter Kontschieder, Vasileios Balntas
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 |
|---|---|---|---|
| Infer scene geometry rather than assume a planar scene (ablation) | OrienterNet · MGL validation | Position R@1m: 14.28 -> 15.78 | Improved |
| Infer scene geometry rather than assume a planar scene (ablation) | OrienterNet · MGL validation | Orientation R@1°: 20.43 -> 22.14 | Improved |
| Replace geometric BEV matching with global map-tile retrieval (ablation) | OrienterNet · MGL validation | Position R@1m: 15.78 -> 2.02 | Hurt |
| Replace geometric BEV matching with featuremetric refinement (ablation) | OrienterNet · MGL validation | Position R@1m: 15.78 -> 08.09 | Hurt |
| Fuse image predictions from a 20-second sequence (ablation) | OrienterNet (MGL) · KITTI Test2 | Lateral R@3m: 88.85 -> 97.44 | Improved |
| Fuse image predictions from a 20-second sequence (ablation) | OrienterNet (MGL) · KITTI Test2 | Longitudinal R@3m: 59.84 -> 95.27 | Improved |
| Fuse image predictions from a 20-second sequence (ablation) | OrienterNet (MGL) · KITTI Test2 | Orientation R@3°: 73.51 -> 97.76 | Improved |
| Pre-train on MGL and fine-tune on KITTI rather than train on KITTI only (ablation) | OrienterNet · KITTI Test2 | Lateral R@3m: 84.77 -> 92.76 | Improved |
| Fuse predictions across multiple frames (ablation) | OrienterNet · Seattle, Aria | Position R@3m: 14.49 -> 61.26 | Improved |
| Fuse predictions across multiple frames (ablation) | OrienterNet · Detroit, Aria | Position R@3m: 32.41 -> 68.77 | Improved |
| Use a larger horizontal field of view (ablation) | OrienterNet · MGL validation | localization accuracy | Improved |
| Remove building and road map elements (ablation) | OrienterNet · MGL validation | localization accuracy | Hurt |
Later papers that compared with this method
3 later accepted papers report OrienterNet as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- RHO: Robust Holistic OSM-Based Metric Cross-View Geo-Localization (CVPR 2026 · better on 3 of 3 shared results)
- HOLO: Homography-Guided Pose Estimator Network for Fine-Grained Visual Localization on SD Maps (CVPR 2026 · better on 2 of 2 shared results)
- VectorReLoc: Reliable Vectorized SD Map Visual Re-localization with Contrastive Feature Alignment (ECCV 2026 · 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
- Lending Orientation to Neural Networks for Cross-View Geo-Localization (CVPR 2019 · similarity 0.88)
- Localizing and Orienting Street Views Using Overhead Imagery (ECCV 2016 · similarity 0.87)
- CrowdDriven: A New Challenging Dataset for Outdoor Visual Localization (ICCV 2021 · similarity 0.87)
- Geometry-Aware Learning of Maps for Camera Localization (CVPR 2018 · similarity 0.87)
- SNAP: Self-Supervised Neural Maps for Visual Positioning and Semantic Understanding (NeurIPS 2023 · similarity 0.87)
- Where Am I and What Will I See: An Auto-Regressive Model for Spatial Localization and View Prediction (ICLR 2025 · similarity 0.85)
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.
