Research map / 3D vision / Depth and stereo
Prior Guided Dropout for Robust Visual Localization in Dynamic Environments
Zhaoyang Huang, Yan Xu, Jianping Shi, Xiaowei Zhou, Hujun Bao, Guofeng Zhang
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 prior guided dropout (ablation) | MapNet · RobotCar Full | position error (m): 44.61 -> 32.64 | Improved |
| add composite self-attention (ablation) | MapNet · RobotCar Full | position error (m): 44.61 -> 30.02 | Improved |
| add composite self-attention (ablation) | PoseNet · RobotCar Full | position error (m): 46.61 -> 62.46 | Hurt |
| add composite self-attention on top of prior guided dropout (ablation) | PoseNet · RobotCar Full | position error (m): 38.56 -> 33.82 | Improved |
| add composite self-attention on top of prior guided dropout (ablation) | MapNet · RobotCar Full | position error (m): 32.64 -> 19.18 | Improved |
| use variational PGO instead of constant PGO (ablation) | AD-MapNet · RobotCar Full | position error (m): 18.84 -> 14.85 | Improved |
| add constant PGO | AD-MapNet · RobotCar Full | orientation error (deg): 4.60 -> 13.73 | Hurt |
| use variational PGO instead of constant PGO (ablation) | AD-PoseNet · RobotCar Loop | position error (m): 7.04 -> 6.49 | Improved |
| add constant PGO | AD-PoseNet · RobotCar Full | position error (m): 33.82 -> 27.35 | Improved |
| use prior guided dropout instead of uniform dropout (T=0.2) (ablation) | AD-MapNet · RobotCar Full | position error (m): 53.69 -> 19.18 | Improved |
| set dropout threshold T=0.2 instead of 0.4 (ablation) | AD-MapNet · RobotCar Full | position error (m): 33.03 -> 19.18 | Improved |
| set dropout threshold T=0.2 instead of 0.1 (ablation) | AD-MapNet · RobotCar Loop | position error (m): 5.65 -> 6.45 | Hurt |
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 3 of 3 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
- AtLoc: Attention Guided Camera Localization (AAAI 2020 · similarity 0.88)
- Long-Term Visual Localization Revisited (TPAMI 2020 · similarity 0.88)
- Uncertainty-Driven 6D Pose Estimation of Objects and Scenes From a Single RGB Image (CVPR 2016 · similarity 0.87)
- Understanding the Limitations of CNN-Based Absolute Camera Pose Regression (CVPR 2019 · similarity 0.87)
- Reloc3r: Large-Scale Training of Relative Camera Pose Regression for Generalizable, Fast, and Accurate Visual Localization (CVPR 2025 · similarity 0.86)
- HyperPose: Hypernetwork-Infused Camera Pose Localization and an Extended Cambridge Landmarks Dataset (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.
