Research map / 3D vision / Depth and stereo
VS-Net: Voting With Segmentation for Visual Localization
Zhaoyang Huang, Han Zhou, Yijin Li, Bangbang Yang, Yan Xu, Xiaowei Zhou, Hujun Bao, Guofeng Zhang, Hongsheng Li
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 scene-specific landmarks with segmentation and voting instead of pixel-wise scene coordinates (ablation) | VS-Net visual localization · 7-Scenes (Avg) | median positional error: 36.5cm -> 2.4cm | Improved |
| use scene-specific landmarks with segmentation and voting instead of pixel-wise scene coordinates (ablation) | VS-Net visual localization · Cambridge Landmarks (Avg) | median positional error: 99cm -> 14cm | Improved |
| add kNN hard-negative mining to prototype triplet segmentation loss (ablation) | VS-Net · 7-Scenes | segmentation accuracy: 0.45 -> 0.80 (Chess); 0.34 -> 0.69 (Fire); 0.25 -> 0.58 (Heads); 0.4 -> 0.75 (Office); 0.43 -> 0.61 (Pumpkin); 0.26 -> 0 | Improved |
| vary the 3D patch size (ablation) | VS-Net · King's College and Chess | median positional error: King's College: 16cm, 15cm, 16cm, 15cm, 17cm, 17cm, 18cm; Chess: 1.54cm, 1.52cm, 1.52cm, 1.52cm, 1.57cm, 1.58cm, 1.65cm | Mixed |
| use VS-Net rather than the listed strongest baselines | visual localization · 7-Scenes (Avg) | median positional error: 2.7cm -> 2.4cm (HSC-Net); 3.5cm -> 2.4cm (DSAC++) | Improved |
| use VS-Net rather than the listed strongest baselines | visual localization · Cambridge Landmarks (Avg) | median positional error: 0.160m -> 0.136m (HSC-Net); 0.194m -> 0.136m (DSAC++) | Improved |
Later papers that compared with this method
2 later accepted papers report VSNet as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- HumanPro: Single-view 3D Clothed Human Reconstruction with Progressive Normal Guidance (AAAI 2026 · better on 3 of 3 shared results)
- OFVL-MS: Once for Visual Localization across Multiple Indoor Scenes (ICCV 2023 · 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
- Yes, we CANN: Constrained Approximate Nearest Neighbors for Local Feature-Based Visual Localization (ICCV 2023 · similarity 0.89)
- Investigating the Role of Image Retrieval for Visual Localization (IJCV 2022 · similarity 0.88)
- RELOCATE: A Simple Training-Free Baseline for Visual Query Localization Using Region-Based Representations (CVPR 2025 · similarity 0.87)
- R-SCoRe: Revisiting Scene Coordinate Regression for Robust Large-Scale Visual Localization (CVPR 2025 · similarity 0.86)
- Enhancing Visual Localization with Cross-Domain Image Generation (ICML 2025 · similarity 0.86)
- GPVK-VL: Geometry-Preserving Virtual Keyframes for Visual Localization under Large Viewpoint Changes (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.
