Research map / 3D vision / Depth and stereo
Are Large-Scale 3D Models Really Necessary for Accurate Visual Localization?
Torsten Sattler, Akihiko Torii, Josef Sivic, Marc Pollefeys, Hajime Taira, Masatoshi Okutomi, Tomas Pajdla
Check 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 spatial re-ranking rather than the nearest-neighbor image pose (ablation) | DenseVLAD · Dubrovnik | median positional error: 3.9 -> 2.9 | Improved |
| Add local SfM pose estimation after spatial re-ranking (ablation) | DenseVLAD · Dubrovnik | median positional error: 2.9 -> 1.0 | Improved |
| Use retrieval followed by local SfM pose estimation (ablation) | DenseVLAD · San Francisco | positional localization accuracy | Improved |
| Use retrieval followed by local SfM pose estimation (ablation) | Disloc · San Francisco | positional localization accuracy | Improved |
| Use Disloc with local SfM rather than a pre-built 3D localization method | Disloc · San Francisco | positional localization accuracy | Improved |
| Use DenseVLAD retrieval with local SfM rather than CPV | DenseVLAD · Dubrovnik | median positional error: 1.0 vs. 0.56 | Hurt |
| Add geometric burstiness re-ranking (ablation) | Disloc · San Francisco | positional localization accuracy | Improved |
| Use NetVLAD instead of DenseVLAD with the nearest-neighbor pose strategy | VLAD retrieval · San Francisco | positional localization accuracy | Hurt |
| Use NetVLAD instead of DenseVLAD with spatial re-ranking and local SfM | VLAD retrieval · San Francisco | positional localization accuracy | No clear effect |
Closest papers by meaning
- Efficient & Effective Prioritized Matching for Large-Scale Image-Based Localization (TPAMI 2016 · similarity 0.91)
- Combining Absolute and Semi-Generalized Relative Poses for Visual Localization (IJCV 2026 · similarity 0.90)
- Investigating the Role of Image Retrieval for Visual Localization (IJCV 2022 · similarity 0.90)
- A Guide to Structureless Visual Localization (IJCV 2026 · similarity 0.90)
- Efficient Global 2D-3D Matching for Camera Localization in a Large-Scale 3D Map (ICCV 2017 · similarity 0.89)
- Enhancing Visual Localization with Cross-Domain Image Generation (ICML 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.
