| Query-Based Adaptive Aggregation for Multi-Dataset Joint Training Toward Universal Visual Place Recognition | | 97.6 | | | QAA-DINOv2-B-8192 | 2025-07-04 |
| A^2GC: Asymmetric Aggregation with Geometric Constraints for Locally Aggregated Descriptors | ✓ Link | 96 | 97.9 | 98.6 | A²GC (DINOv2-B, 588x588) | 2025-11-18 |
| Pair-VPR: Place-Aware Pre-training and Contrastive Pair Classification for Visual Place Recognition with Vision Transformers | ✓ Link | 95.4 | 97.3 | 97.7 | Pair-VPR-p | 2024-10-09 |
| SelaVPR++: Towards Seamless Adaptation of Foundation Models for Efficient Place Recognition | ✓ Link | 94.9 | | | SelaVPR++ (DINOv2-B, BoQ12288D, unified training) | 2025-02-23 |
| Lost in the Tail: Addressing Geographic Imbalance in Urban Visual Place Recognition | | 94.7 | | | CliqueMining + LB loss | 2026-06-30 |
| SuperPlace: The Renaissance of Classical Feature Aggregation for Visual Place Recognition in the Era of Foundation Models | ✓ Link | 94.6 | | | SuperPlace NVL-FT2 (DINOv2-B, GPMS, 8192D, 224px) | 2025-06-16 |
| Towards Implicit Aggregation: Robust Image Representation for Place Recognition in the Transformer Era | ✓ Link | 94.5 | 97.3 | 98 | ImAge (6144D, unified dataset) | 2025-11-08 |
| "Close, But Not There: Boosting Geographic Distance Sensitivity in Visual Place Recognition" | ✓ Link | 94.2 | 97.2 | 97.4 | DINOv2 SALAD + CliqueMining | 2024-07-02 |
| SAGE: Spatial-visual Adaptive Graph Exploration for Efficient Visual Place Recognition | ✓ Link | 94.2 | 97.8 | 98.1 | SAGE ViT-L (without cross-image encoder) | 2025-09-30 |
| Region Matters: Efficient and Reliable Region-Aware Visual Place Recognition | ✓ Link | 94.1 | | | FoL++ (ViT-L,reranking,322px) | 2026-04-24 |
| BoQ: A Place is Worth a Bag of Learnable Queries | ✓ Link | 93.8 | 96.8 | 97 | BoQ | 2024-05-12 |
| Pair-VPR: Place-Aware Pre-training and Contrastive Pair Classification for Visual Place Recognition with Vision Transformers | ✓ Link | 93.7 | 97.2 | 97.3 | Pair-VPR-s | 2024-10-09 |
| Focus on Local: Finding Reliable Discriminative Regions for Visual Place Recognition | ✓ Link | 93.5 | 96.9 | 97.6 | FoL | 2025-04-14 |
| UniPR-3D: Towards Universal Visual Place Recognition with Visual Geometry Grounded Transformer | ✓ Link | 93.2 | | 97.9 | UniPR-3D* (MegaLoc training setting, single frame) | 2025-12-24 |
| Focus on Local: Finding Reliable Discriminative Regions for Visual Place Recognition | ✓ Link | 93.1 | 96.9 | 97.4 | FoL-global | 2025-04-14 |
| EffoVPR: Effective Foundation Model Utilization for Visual Place Recognition | | 92.8 | 97.2 | 97.4 | EffoVPR | 2024-05-28 |
| EfficientVPR: Toward Efficient Visual Place Recognition via Scene-Aware Prompt Tuning and Adaptive Feature Enhancement | ✓ Link | 92.8 | 96.6 | | EfficientVPR (DINOv2-S, 322×322 inference) | 2026-06-01 |
| EDTformer: An Efficient Decoder Transformer for Visual Place Recognition | ✓ Link | 92.7 | | | EDTformer (4096D, LoPA rank24) | 2024-12-01 |
| Faster or Stronger: Towards Flexible Visual Place Recognition via Weighted Aggregation and Token Pruning | | 92.7 | | | WeiAD (DINOv2 ViT-L) | 2026-05-19 |
| VLAD-BuFF: Burst-aware Fast Feature Aggregation for Visual Place Recognition | ✓ Link | 92.6 | | | VLAD-BuFF (384-PrePool, PCA whitening,4096D) | 2024-09-28 |
| SuperVLAD: Compact and Robust Image Descriptors for Visual Place Recognition | ✓ Link | 92.4 | 96.4 | 96.9 | SuperVLAD (DINOv2-B, 4 clusters, no ghost cluster, no cross-image encoder) | 2024-09-26 |
| Optimal Transport Aggregation for Visual Place Recognition | ✓ Link | 92.2 | 96.4 | 97 | DINOv2 SALAD | 2023-11-27 |
| BoQ: A Place is Worth a Bag of Learnable Queries | ✓ Link | 91.2 | 95.3 | 96.1 | BoQ (ResNet-50) | 2024-05-12 |
| D²-VPR: A Parameter-efficient Visual-foundation-model-based Visual Place Recognition Method via Knowledge Distillation and Deformable Aggregation | ✓ Link | 91.1 | 96.1 | | D2-VPR (singleton encoder,10752D) | 2025-11-16 |
| MegaLoc: One Retrieval to Place Them All | ✓ Link | 91 | | 95.8 | MegaLoc (8448D) | 2025-02-24 |
| Towards Seamless Adaptation of Pre-trained Models for Visual Place Recognition | ✓ Link | 90.8 | 96.4 | 97.2 | SelaVPR | 2024-02-22 |
| $R^{2}$Former: Unified $R$etrieval and $R$eranking Transformer for Place Recognition | ✓ Link | 90 | 95.1 | 96.9 | R2Former (ViT-Base, MSLS-trained, top-100 reranking) | 2023-04-06 |
| EigenPlaces: Training Viewpoint Robust Models for Visual Place Recognition | ✓ Link | 89.5 | | | EigenPlaces (ResNet-50, 512D) | 2023-08-21 |
| CricaVPR: Cross-image Correlation-aware Representation Learning for Visual Place Recognition | ✓ Link | 88.5 | 95.1 | 95.7 | CricaVPR (inference batch1,10752D) | 2024-02-29 |
| MixVPR: Feature Mixing for Visual Place Recognition | ✓ Link | 88.2 | 93.1 | 94.3 | MixVPR | 2023-03-03 |
| Rethinking Visual Geo-localization for Large-Scale Applications | ✓ Link | 86.7 | 92.1 | 93.4 | CosPlace (ResNet-101 2048-D) | 2022-04-05 |
| Deep Homography Estimation for Visual Place Recognition | ✓ Link | 84.1 | 91.4 | 93.1 | DHE-VPR (jointly fine-tuned) | 2024-02-25 |
| GSV-Cities: Toward Appropriate Supervised Visual Place Recognition | ✓ Link | 83.4 | 90.5 | 92.3 | Conv-AP | 2022-10-19 |
| Structured Pruning for Efficient Visual Place Recognition | | 82.7 | | | NetVLAD (structured pruning40%,gamma0.23) | 2024-09-12 |
| Global Proxy-based Hard Mining for Visual Place Recognition | ✓ Link | 82.0 | 90.4 | 91.4 | NetVLAD (with GPM) | 2023-02-28 |
| Generalized Contrastive Optimization of Siamese Networks for Place Recognition | ✓ Link | 80.9 | 90.7 | 92.6 | ResNeXt GCL + PCA | 2021-03-11 |
| Patch-NetVLAD: Multi-Scale Fusion of Locally-Global Descriptors for Place Recognition | ✓ Link | 79.5 | 86.2 | 87.7 | Patch-NetVLAD | 2021-03-02 |
| Generalized Contrastive Optimization of Siamese Networks for Place Recognition | ✓ Link | 75.5 | 86.1 | 88.5 | ResNeXt GCL | 2021-03-11 |
| Rethinking Visual Geo-localization for Large-Scale Applications | ✓ Link | | 89.9 | 91.8 | CosPlace | 2022-04-05 |
| EigenPlaces: Training Viewpoint Robust Models for Visual Place Recognition | ✓ Link | | 93.8 | 95 | EigenPlaces (ResNet-50, 2048D) | 2023-08-21 |
| UniPR-3D: Towards Universal Visual Place Recognition with Visual Geometry Grounded Transformer | ✓ Link | | 96 | | UniPR-3D (GSV-Cities single-frame release) | 2025-12-24 |
| VLAD-BuFF: Burst-aware Fast Feature Aggregation for Visual Place Recognition | ✓ Link | | 96.2 | | VLAD-BuFF (PrePool256,PCA whitening4096D) | 2024-09-28 |
| VLAD-BuFF: Burst-aware Fast Feature Aggregation for Visual Place Recognition | ✓ Link | | | 96.9 | VLAD-BuFF (PCA whitening,8192D) | 2024-09-28 |
| Lost in the Tail: Addressing Geographic Imbalance in Urban Visual Place Recognition | | | 96.5 | 97 | SALAD + LB loss | 2026-06-30 |
| Region Matters: Efficient and Reliable Region-Aware Visual Place Recognition | ✓ Link | | 98 | 98.2 | FoL++ (ViT-L,global,322px) | 2026-04-24 |
| D²-VPR: A Parameter-efficient Visual-foundation-model-based Visual Place Recognition Method via Knowledge Distillation and Deformable Aggregation | ✓ Link | | | 96.6 | D2-VPR (singleton encoder) | 2025-11-16 |
| EDTformer: An Efficient Decoder Transformer for Visual Place Recognition | ✓ Link | | 96.9 | | EDTformer (4096D, 4decoder blocks) | 2024-12-01 |
| EDTformer: An Efficient Decoder Transformer for Visual Place Recognition | ✓ Link | | | 97.2 | EDTformer (4096D, 2decoder blocks, 64learned tokens, LoPA rank4) | 2024-12-01 |
| Faster or Stronger: Towards Flexible Visual Place Recognition via Weighted Aggregation and Token Pruning | | | 96.8 | | WeiAD + WeiToP (DINOv2 ViT-L, retention 0.9) | 2026-05-19 |
| SuperPlace: The Renaissance of Classical Feature Aggregation for Visual Place Recognition in the Era of Foundation Models | ✓ Link | | 97.8 | 98.1 | SuperPlace SP-NVL-FT2 (DINOv2-L, GPMS, 8192D, 322px) | 2025-06-16 |
| SelaVPR++: Towards Seamless Adaptation of Foundation Models for Efficient Place Recognition | ✓ Link | | 98 | 98.2 | SelaVPR++ (DINOv2-L, binary512D + float4096D, rerank100) | 2025-02-23 |