Research map / Detection and segmentation / CNN features
StructVPR++: Distill Structural and Semantic Knowledge With Weighting Samples for Visual Place Recognition
Yanqing Shen, Sanping Zhou, Jingwen Fu, Ruotong Wang, Shitao Chen, Nanning Zheng
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 |
|---|---|---|---|
| replace PSPNet with UPerNet segmentation (ablation) | StructVPR++ · MSLS val | Recall@1: 84.3 -> 83.5 | Hurt |
| replace PSPNet with HRNetV2 segmentation (ablation) | StructVPR++ · MSLS val | Recall@1: 84.3 -> 82.6 | Hurt |
| replace PSPNet with HRNetV2 segmentation (ablation) | StructVPR++ · MSLS test | Recall@1: 65.7 -> 65.8 | No clear effect |
| replace PSPNet with UPerNet segmentation (ablation) | StructVPR++ · Nordland | Recall@1: 58.4 -> 58.2 | No clear effect |
| replace PSPNet with HRNetV2 segmentation (ablation) | StructVPR++ · Nordland | Recall@10: 84.1 -> 84.7 | No clear effect |
| replace PSPNet with UPerNet segmentation (ablation) | Seg-branch · MSLS test | Recall@10: 70.1 -> 70.2 | No clear effect |
| cluster segmentation into 150 classes instead of 5 or 6 (ablation) | seg-branch · MSLS val | Recall | Hurt |
| cluster into only 3 classes instead of 5 or 6 (ablation) | seg-branch · MSLS val | Recall | Hurt |
| ignore dynamic objects (C=5 instead of C=6) (ablation) | seg-branch · MSLS val | Recall | Improved |
| add appropriate prior class weights (ablation) | seg-branch · MSLS val | Recall | Improved |
| vary group partition threshold Nt (3 to 20) (ablation) | StructVPR · MSLS val | Recall | No clear effect |
Closest papers by meaning
- DistilVPR: Cross-Modal Knowledge Distillation for Visual Place Recognition (AAAI 2024 · similarity 0.88)
- EffoVPR: Effective Foundation Model Utilization for Visual Place Recognition (ICLR 2025 · similarity 0.88)
- Focus on Local: Finding Reliable Discriminative Regions for Visual Place Recognition (AAAI 2025 · similarity 0.87)
- Towards Implicit Aggregation: Robust Image Representation for Place Recognition in the Transformer Era (NeurIPS 2025 · similarity 0.87)
- MutualVPR: A Mutual Learning Framework for Resolving Supervision Inconsistencies via Adaptive Clustering (NeurIPS 2025 · similarity 0.87)
- GCD: Advancing Vision-Language Models for Incremental Object Detection via Global Alignment and Correspondence Distillation (AAAI 2025 · 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.
