Research map / Label-efficient and robust learning / Knowledge distillation
D²-VPR: A Parameter-efficient Visual-foundation-model-based Visual Place Recognition Method via Knowledge Distillation and Deformable Aggregation
Zheyuan Zhang, Jiwei Zhang, Boyu Zhou, Linzhimeng Duan, Hong Chen
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 knowledge distillation from CricaVPR teacher before training (ablation) | D2-VPR (DINOv2-small) · MSLS-val | Recall@1: 85.1 -> 88.4 | Improved |
| add fine-tuning after distillation (ablation) | D2-VPR (DINOv2-small) · MSLS-val | Recall@1: 88.4 -> 91.6 | Improved |
| add distillation recovery module (DRM) (ablation) | D2-VPR (DINOv2-small) · MSLS-val | Recall@1: 89.9 -> 90.5 | Improved |
| add top-down deformable aggregator (TDDA) (ablation) | D2-VPR (DINOv2-small) · MSLS-val | Recall@1: 90.5 -> 91.6 | Improved |
| add TDDA (ablation) | D2-VPR (DINOv2-small) · MSLS-val | Recall@10: 96.2 -> 97.2 | Improved |
| train DINOv2-small directly with CricaVPR pipeline instead of DINOv2-base (ablation) | CricaVPR · MSLS-val | Recall | Hurt |
| add cross-image encoder to D2-VPR | D2-VPR · Pitts30k (224x224) | Recall@1: 91.7 -> 94.9 | Improved |
| add cross-image encoder to D2-VPR | D2-VPR · Nordland (224x224) | Recall@1: 77.1 -> 86.6 | Improved |
| use deformable aggregator with sampling operations | D2-VPR · AmsterTime inference time | inference speed | No clear effect |
Closest papers by meaning
- EffoVPR: Effective Foundation Model Utilization for Visual Place Recognition (ICLR 2025 · similarity 0.92)
- Towards Implicit Aggregation: Robust Image Representation for Place Recognition in the Transformer Era (NeurIPS 2025 · similarity 0.91)
- Optimal Transport Aggregation for Visual Place Recognition (CVPR 2024 · similarity 0.91)
- DistilVPR: Cross-Modal Knowledge Distillation for Visual Place Recognition (AAAI 2024 · similarity 0.91)
- Focus on Local: Finding Reliable Discriminative Regions for Visual Place Recognition (AAAI 2025 · similarity 0.91)
- Towards Seamless Adaptation of Pre-trained Models for Visual Place Recognition (ICLR 2024 · similarity 0.90)
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.
