Research map / 3D vision / Depth and stereo
Beyond Cross-View Image Retrieval: Highly Accurate Vehicle Localization Using Satellite Image
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 iterative pose optimization with satellite-to-ground feature projection instead of fine-grained image retrieval | Vehicle localization · KITTI Test1 | lateral accuracy at d=1: 35.54 vs 18.61 | Helped |
| Use iterative pose optimization with satellite-to-ground feature projection instead of fine-grained image retrieval | Vehicle localization · Ford Log1 | lateral accuracy at d=1: 46.10 vs 20.33 | Helped |
| Add satellite-to-ground projection rather than ground-to-satellite projection (ablation) | Pose optimization with geometry projection · KITTI Test1 | lateral accuracy at d=1: 27.72 -> 35.54 | Helped |
| Use homography instead of polar transform (ablation) | Satellite-to-ground projection · KITTI Test1 | lateral accuracy at d=1: 18.98 -> 35.54 | Helped |
| Replace implicit network mapping with explicit geometry projection (ablation) | Ground-to-satellite projection · KITTI Test1 | lateral accuracy at d=1: 20.30 -> 27.72 | Helped |
| Use LM instead of SGD for pose optimization (ablation) | Cross-view pose optimization · KITTI Test1 | lateral accuracy at d=1: 16.86 -> 35.54 | Helped |
| Use LM instead of Adam for pose optimization (ablation) | Cross-view pose optimization · KITTI Test1 | lateral accuracy at d=1: 7.13 -> 35.54 | Helped |
| Use LM instead of a learned network optimizer (ablation) | Cross-view pose optimization · KITTI Test1 | lateral accuracy at d=1: 27.14 -> 35.54 | Helped |
| Add longitudinal pose supervision (ablation) | Proposed method · KITTI Test1 | lateral accuracy at d=1: 25.63 -> 35.54 | Helped |
| Use iterative coarse-to-fine updates instead of completing coarse-level iterations before advancing (ablation) | Multi-level LM optimization · KITTI Test1 | lateral accuracy at d=1: 23.32 -> 35.54 | Helped |
| Filter inaccurate GPS-tagged training samples (ablation) | Proposed method · Ford Log1 | lateral accuracy at d=1: 26.67 -> 46.10 | Helped |
| Add a confidence map to weight salient features (ablation) | Feature matching objective · Different test sets | localization performance | No clear effect |
Closest papers by meaning
- PIDLoc: Cross-View Pose Optimization Network Inspired by PID Controllers (CVPR 2025 · similarity 0.89)
- View From Above: Orthogonal-View aware Cross-view Localization (CVPR 2024 · similarity 0.88)
- BevSplat: Resolving Height Ambiguity via Feature-Based Gaussian Primitives for Weakly-Supervised Cross-View Localization (NeurIPS 2025 · similarity 0.88)
- From Coarse to Fine: A Matching and Alignment Framework for Unsupervised Cross-View Geo-Localization (AAAI 2025 · similarity 0.88)
- Aligning Geometric Spatial Layout in Cross-View Geo-Localization via Feature Recombination (AAAI 2024 · similarity 0.87)
- Weather-Robust Cross-View Geo-Localization via Prototype-Based Semantic Part Discovery (NeurIPS 2026 · similarity 0.87)
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.