Research map / 3D vision / Human pose estimation
Local Supports Global: Deep Camera Relocalization With Sequence Enhancement
Fei Xue, Xin Wang, Zike Yan, Qiuyuan Wang, Junqiu Wang, Hongbin Zha
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 content augmentation to the basic model (ablation) | Ours (basic) · Oxford RobotCar (average) | mean translation error (m): 64.58 -> 41.46 | Improved |
| add content augmentation to the basic model (ablation) | Ours (basic) · Oxford RobotCar (average) | mean rotation error (deg): 13.87 -> 7.88 | Improved |
| add motion-based refinement (pose graph) to content-augmented model (ablation) | Ours (w/ content) · Oxford RobotCar (average) | mean translation error (m): 41.46 -> 25.84 | Improved |
| add motion-based refinement (pose graph) to content-augmented model (ablation) | Ours (w/ content) · Oxford RobotCar (average) | mean rotation error (deg): 7.88 -> 4.99 | Improved |
| use sequences with VO component instead of single images (vs PoseNet17) | Ours · 7-Scenes average | median translation error (m): 0.23 -> 0.19 | Improved |
| use content augmentation and pose graph instead of MapNet's training-time motion constraints | Ours · Oxford RobotCar average | mean translation error (m): 29.83 -> 25.84 | Improved |
| use sequence model instead of MapNet on regular small indoor scene | Ours · 7-Scenes Chess | median translation error (m): 0.08 -> 0.09 | No clear effect |
| use sequence model instead of MapNet on night sequences | Ours · Oxford RobotCar FULL8 | mean translation error (m): 430.49 -> 430.54 | No clear effect |
| use sequence model vs MapNet under varied weather and day-night | Ours · Oxford RobotCar FULL3-FULL8 average | mean translation error (m): 227.69 -> 193.98 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- MapNet — MapNet: An Allocentric Spatial Memory for Mapping Environments (CVPR 2018) · better on 8 of 8 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
- Accelerated Coordinate Encoding: Learning to Relocalize in Minutes Using RGB and Poses (CVPR 2023 · similarity 0.89)
- Map-Free Visual Relocalization: Metric Pose Relative to a Single Image (ECCV 2022 · similarity 0.88)
- Long-Term Visual Localization Revisited (TPAMI 2020 · similarity 0.88)
- Reloc3r: Large-Scale Training of Relative Camera Pose Regression for Generalizable, Fast, and Accurate Visual Localization (CVPR 2025 · similarity 0.87)
- Change-Robust Online Topological Memory for Long-Term Relocalization and Semantic Navigation (NeurIPS 2026 · similarity 0.86)
- Leveraging Consistent Spatio-Temporal Correspondence for Robust Visual Odometry (AAAI 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.
