Research map / Label-efficient and robust learning / Knowledge distillation
StructVPR: Distill Structural Knowledge With Weighting Samples for Visual Place Recognition
Yanqing Shen, Sanping Zhou, Jingwen Fu, Ruotong Wang, Shitao Chen, Nanning Zheng
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 |
|---|---|---|---|
| Distill segmentation-branch structural features into RGB features | StructVPR · MSLS val | Recall@1: 75.8 -> 83.0 | Improved |
| Distill segmentation-branch structural features into RGB features | StructVPR · MSLS challenge | Recall@1: 55.1 -> 64.5 | Improved |
| Add SuperPoint feature matching and SuperGlue re-ranking | StructVPR · Nordland test | Recall@1: 56.1 -> 83.5 | Improved |
| Use all samples for distillation instead of selectively choosing samples (ablation) | StructVPR · MSLS val | Recall@1: 83.0 -> 78.4 | Hurt |
| Distill only from group S2 (D4) (ablation) | StructVPR · MSLS val | Recall@1: 75.8 -> 71.6 | Hurt |
| Distill only from group R2 (ablation) | StructVPR · MSLS val | Recall@1: 75.8 -> 73.8 | Hurt |
| Use a two-branch partition instead of teacher-only partition, with per-sample weighting (ablation) | StructVPR · MSLS val | Recall@1: 79.7 -> 83.0 | Improved |
| Replace constant group weights with per-sample weights (ablation) | GP-D · MSLS val | Recall@1: 82.2 -> 83.0 | Improved |
| Use three semantic classes instead of six or 150 (ablation) | Seg-branch SLME · MSLS | Recall | Mixed |
| Introduce appropriate prior weights for semantic classes (ablation) | Seg-branch SLME · MSLS | Recall | Improved |
Closest papers by meaning
- EffoVPR: Effective Foundation Model Utilization for Visual Place Recognition (ICLR 2025 · similarity 0.89)
- DistilVPR: Cross-Modal Knowledge Distillation for Visual Place Recognition (AAAI 2024 · similarity 0.88)
- Focus on Local: Finding Reliable Discriminative Regions for Visual Place Recognition (AAAI 2025 · similarity 0.88)
- Towards Implicit Aggregation: Robust Image Representation for Place Recognition in the Transformer Era (NeurIPS 2025 · similarity 0.88)
- Optimal Transport Aggregation for Visual Place Recognition (CVPR 2024 · similarity 0.87)
- CricaVPR: Cross-image Correlation-aware Representation Learning for Visual Place Recognition (CVPR 2024 · 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.
