Research map / Vision-language and multimodal / Multimodal LLMs
Efficient Visual Place Recognition Through Multimodal Semantic Knowledge Integration
Sitao Zhang, Hongda Mao, Qingshuang Chen, Yelin Kim
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 semantic-aware aggregation (ablation) | SemVPR · Pitts30k (64-dim) | Recall@1: 77.4 -> 90.1 | Improved |
| add semantic-aware aggregation (ablation) | SemVPR · Pitts30k (1024-dim) | Recall@1: 91.7 -> 94.2 | Improved |
| add local semantics alignment of teacher (ablation) | SemVPR · Pitts30k (64-dim) | Recall@1: 80.3 -> 90.1 | Improved |
| add local semantics alignment of teacher (ablation) | SemVPR · Pitts30k (1024-dim) | Recall@1: 93.5 -> 94.2 | Improved |
| add nested descriptor learning (ablation) | SemVPR · Pitts30k (64-dim) | Recall@1: 82.6 -> 90.1 | Improved |
| add nested descriptor learning (ablation) | SemVPR · Pitts30k (256-dim) | Recall@1: 93.5 vs w/o NDL 91.4 | Improved |
| add nested descriptor learning (ablation) | SemVPR · Pitts30k (1024-dim) | Recall@1: 94.4 -> 94.2 | Hurt |
| fine-tune DINOv2 with SemVPR instead of frozen DINOv2 | DINOv2 CLS · Pitts30k (64-dim) | Recall@1: 67.6 -> 90.1 | Improved |
| use SemVPR 1024-dim instead of SALAD 8448-dim | VPR descriptor · Pitts30k | Recall@1: 92.2 -> 94.2 | Improved |
| use SemVPR 1024-dim instead of SALAD | VPR descriptor · SF-XL Night | Recall@1: 46.6 -> 56.2 | Improved |
| use SemVPR global descriptor instead of SelaVPR two-stage re-ranking | VPR · Tokyo24/7 | Recall@1: 94.0 -> 96.8 | Improved |
| use SemVPR 1024-dim instead of SALAD | VPR descriptor · MSLS-val | Recall@1: 91.2 -> 89.7 | Hurt |
Closest papers by meaning
- DistilVPR: Cross-Modal Knowledge Distillation for Visual Place Recognition (AAAI 2024 · similarity 0.89)
- Towards Implicit Aggregation: Robust Image Representation for Place Recognition in the Transformer Era (NeurIPS 2025 · similarity 0.88)
- Towards Seamless Adaptation of Pre-trained Models for Visual Place Recognition (ICLR 2024 · similarity 0.88)
- MutualVPR: A Mutual Learning Framework for Resolving Supervision Inconsistencies via Adaptive Clustering (NeurIPS 2025 · similarity 0.87)
- EffoVPR: Effective Foundation Model Utilization for Visual Place Recognition (ICLR 2025 · similarity 0.87)
- Optimal Transport Aggregation 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.
