Visual Place Recognition on Pitts30k
95.4Recall@1Pair-VPR-p
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Pair-VPR-pEnc. Size=ViT-G2024.10 | 95.4 | 97.5 | 98 | — | — | — | |
| CricaVPRtwo-stage=false2024.05 | 94.9 | 97.3 | 98.2 | — | — | — | |
| CricaVPR2025.04 | 94.9 | 97.3 | 98.2 | — | — | — | |
| Pair-VPR-sEnc. Size=ViT-B2024.10 | 94.7 | 97.2 | 97.8 | — | — | — | |
| FoLmode=re-ranking2025.04 | 94.5 | 97.4 | 98.2 | — | — | — | |
| QAABackbones=DINOv2-B, Cd=81922025.07 | 94.4 | — | — | — | — | — | |
| QAABackbones=DINOv2-B, Cd=20482025.07 | 94 | — | — | — | — | — | |
| QAABackbones=DINOv2-B, Cd=10242025.07 | 93.9 | — | — | — | — | — | |
| EffoVPR-Rtwo-stage=true2024.05 | 93.9 | 97.4 | 98.5 | — | — | — | |
| FoLmode=global2025.04 | 93.9 | 97.2 | 98.1 | — | — | — | |
| QAABackbones=DINOv2-B, Cd=40962025.07 | 93.8 | — | — | — | — | — | |
| BoQBackbones=DINOv2-B, Cd=122882025.07 | 93.7 | — | — | — | — | — | |
| La-CricaVPRDim (d)=+1024, Fusion=ADS+LLP, Text Backbone=BGE-L2026.02 | 93.3 | 96.7 | 97.7 | — | — | — | |
| CricaVPRDim (d)=107522026.02 | 93.2 | 96.7 | 97.7 | — | — | — | |
| CricaVPRDim=10752, Language Augmentation=false2026.02 | 93.2 | 96.7 | 97.7 | — | — | — | |
| SegVLAD-FineTVariant=M2024.09 | 93.1 | 96.8 | — | — | — | — | |
| ProGeo (Transformer)Descriptor dimension=512, Train set=SF-XL2024.06 | 93 | 98.3 | — | — | — | — | |
| SegVLAD-FineTVariant=D2024.09 | 92.9 | 96.8 | — | — | — | — | |
| BoQ*Backbone=DINOv2, Dim.=8192, Retrained with L_lb=true2026.06 | 92.9 | 96.6 | 97.5 | — | — | — | |
| SelaVPRtwo-stage=true2024.05 | 92.8 | 96.8 | 97.7 | — | — | — | |
| SelaVPREnc. Size=ViT-L2024.10 | 92.8 | 96.8 | 97.7 | — | — | — | |
| SelaVPR2025.04 | 92.8 | 96.8 | 97.7 | — | — | — | |
| SALAD*Backbone=DINOv2, Dim.=8448, Retrained with L_lb=true2026.06 | 92.7 | 96.8 | 97.7 | — | — | — | |
| SALAD CMBackbones=DINOv2-B, Cd=84482025.07 | 92.6 | — | — | — | — | — | |
| EigenPlaces2024.09 | 92.6 | 96.7 | — | — | — | — | |
| SALAD2024.09 | 92.6 | 96.5 | — | — | — | — | |
| BoQBackbone=DINOv2, Dim.=8192, Retrained with L_lb=false2026.06 | 92.6 | 96.4 | 97.4 | — | — | — | |
| EigenPlaceBackbones=ResNet-50, Cd=20482025.07 | 92.5 | — | — | — | — | — | |
| EigenPlacesBackbone=ResNet-50, Descriptor Dimension=20482023.08 | 92.5 | — | — | — | — | — | |
| EigenPlacesBackbone=ResNet-502024.05 | 92.5 | — | — | — | — | — | |
| EigenPlacestwo-stage=false2024.05 | 92.5 | 96.8 | 97.6 | — | — | — | |
| BoQtwo-stage=false2024.05 | 92.5 | 96.4 | 97.4 | — | — | — | |
| EigenPlacesEnc. Size=ResNet-502024.10 | 92.5 | 96.8 | 97.6 | — | — | — | |
| EigenPlaces2025.04 | 92.5 | 96.8 | 97.6 | — | — | — | |
| BoQExtrac. (ms)=7, Rerank. (ms)=02024.05 | 92.4 | 96.2 | — | — | — | — | |
| BoQBackbone=ResNet-502024.05 | 92.4 | — | — | — | — | — | |
| SALADtwo-stage=false2024.05 | 92.4 | 96.3 | 97.4 | — | — | — | |
| BoQEnc. Size=ResNet-502024.10 | 92.4 | 96.5 | 97.4 | — | — | — | |
| SALADEnc. Size=ViT-B2024.10 | 92.4 | 96.3 | 97.4 | — | — | — | |
| SALAD2025.04 | 92.4 | 96.3 | 97.4 | — | — | — | |
| La-SALADDim (d)=+1024, Fusion=ADS+LLP, Text Backbone=BGE-L2026.02 | 92.3 | 96.3 | 97.4 | — | — | — | |
| SALADBackbone=DINOv2, Dim.=8448, Retrained with L_lb=false2026.06 | 92.3 | 96.3 | 97.3 | — | — | — | |
| SALADDim (d)=84482026.02 | 92.2 | 96.1 | 97.4 | — | — | — | |
| SALADDim=8448, Language Augmentation=false2026.02 | 92.2 | 96.1 | 97.4 | — | — | — | |
| EigenPlacesBackbone=ResNet-50, Descriptor Dimension=5122023.08 | 91.9 | — | — | — | — | — | |
| La-MixVPRDim (d)=+1024, Fusion=ADS+LLP, Text Backbone=BGE-L2026.02 | 91.9 | 95.7 | 96.7 | — | — | — | |
| ProGeo (CNN)Descriptor dimension=512, Train set=SF-XL2024.06 | 91.8 | 97.4 | — | — | — | — | |
| La-CricaVPRDim=+1024, Language Augmentation=true, Fusion=CAT Fusion2026.02 | 91.8 | 96.3 | 97.5 | — | — | — | |
| MixVPRDim (d)=40962026.02 | 91.6 | 95.6 | 96.4 | — | — | — | |
| MixVPRDim=4096, Language Augmentation=false2026.02 | 91.6 | 95.6 | 96.4 | — | — | — | |
| MixVPRBackbones=ResNet-50, Cd=40962025.07 | 91.5 | — | — | — | — | — | |
| MixVPRBackbone=ResNet-50, Descriptor Dimension=40962023.08 | 91.5 | — | — | — | — | — | |
| MixVPRBackbone=ResNet-502024.05 | 91.5 | — | — | — | — | — | |
| MixVPRtwo-stage=false2024.05 | 91.5 | 95.5 | 96.4 | — | — | — | |
| MixVPREnc. Size=ResNet-502024.10 | 91.5 | 95.5 | 96.3 | — | — | — | |
| La-SALADDim=+1024, Language Augmentation=true, Fusion=CAT Fusion2026.02 | 91.5 | 96.1 | 97.1 | — | — | — | |
| MixVPRLearning Protocol=Supervised Learning2026.01 | 91.5 | 95.5 | — | — | — | — | |
| MixVPR2024.09 | 91.5 | 95.5 | — | — | — | — | |
| R2 FormerDescriptor dimension=256, Train set=MSLS2024.06 | 91.1 | 95.2 | — | — | — | — | |
| R2FormerExtrac. (ms)=31, Rerank. (ms)=4002024.05 | 91.1 | 95.2 | — | — | — | — | |
| R2Formertwo-stage=true2024.05 | 91.1 | 95.2 | 96.3 | — | — | — | |
| R2FormerEnc. Size=ViT-S2024.10 | 91.1 | 95.2 | 96.3 | — | — | — | |
| CosPlaceBackbones=ResNet-50, Cd=20482025.07 | 90.9 | — | — | — | — | — | |
| CosPlaceBackbone=ResNet-50, Descriptor Dimension=20482023.08 | 90.9 | — | — | — | — | — | |
| CosPlaceBackbone=ResNet-502024.05 | 90.9 | — | — | — | — | — | |
| CosPlaceEnc. Size=ResNet-502024.10 | 90.9 | 95.7 | 96.7 | — | — | — | |
| La-EigenPlaceDim=+1024, Language Augmentation=true, Fusion=CAT Fusion2026.02 | 90.8 | 95.7 | 96.8 | — | — | — | |
| La-MixVPRDim=+1024, Language Augmentation=true, Fusion=CAT Fusion, Base Backbone=MixVPR-40962026.02 | 90.8 | 95.2 | 96.4 | — | — | — | |
| EigenPlacesDim=2048, Language Augmentation=false2026.02 | 90.6 | 95.4 | 96.8 | — | — | — | |
| MixVPRDim=512, Language Augmentation=false2026.02 | 90.6 | 95.6 | 96.3 | — | — | — | |
| Conv-APBackbones=ResNet-50, Cd=40962025.07 | 90.5 | — | — | — | — | — | |
| Conv-APBackbone=ResNet-50, Descriptor Dimension=40962023.08 | 90.5 | — | — | — | — | — | |
| Conv-APBackbone=ResNet-50, Descriptor Dimension=81922023.08 | 90.5 | — | — | — | — | — | |
| Conv-APBackbone=ResNet-502024.05 | 90.5 | — | — | — | — | — | |
| MixVPRBackbone=ResNet-50, Descriptor Dimension=5122023.08 | 90.4 | — | — | — | — | — | |
| Conv-APBackbone=ResNet-50, Descriptor Dimension=20482023.08 | 90.4 | — | — | — | — | — | |
| MixVPRDescriptor dimension=512, Train set=GSV-Cities2024.06 | 90.4 | — | — | — | — | — | |
| PaVPR baselineEnc. Size=ViT-B2024.10 | 90.4 | 95.4 | 96.8 | — | — | — | |
| CosPlaceLearning Protocol=Supervised Learning2026.01 | 90.4 | 95.7 | — | — | — | — | |
| CosPlace2024.09 | 90.4 | 95.7 | — | — | — | — | |
| CosPlaceBackbone=ResNet-50, Descriptor Dimension=5122023.08 | 90.2 | — | — | — | — | — | |
| STUNBackbone=Fine-tuned2026.05 | 90 | 95.2 | 96.5 | 0.423 | 0.475 | 0.487 | |
| DELGRe-ranking=true, Training Dataset=MSLS2023.03 | 89.9 | 95.4 | 96.7 | — | — | — | |
| DELGExtrac. (ms)=190, Rerank. (ms)=352002024.05 | 89.9 | 95.4 | — | — | — | — | |
| La-MixVPRDim=+1024, Language Augmentation=true, Fusion=CAT Fusion, Base Backbone=MixVPR-5122026.02 | 89.9 | 95.3 | 96.4 | — | — | — | |
| EigenPlacesBackbone=VGG-16, Descriptor Dimension=5122023.08 | 89.7 | — | — | — | — | — | |
| EigenPlacesBackbone=VGG-162024.05 | 89.7 | — | — | — | — | — | |
| EigenPlacesBackbone=ResNet-50, Descriptor Dimension=1282023.08 | 89.6 | — | — | — | — | — | |
| EffoVPR-ZSZero-Shot=true, re-ranking=top-100 global candidates2024.05 | 89.4 | — | — | — | — | — | |
| SFRStwo-stage=false2024.05 | 89.4 | 94.7 | 95.9 | — | — | — | |
| L2 (Baseline)Backbone=Frozen2026.05 | 89.4 | 95.3 | 96.6 | 0.516 | 0.576 | 0.588 | |
| PA-ScoreBackbone=Frozen2026.05 | 89.4 | 95.3 | 96.6 | 0.765 | 0.825 | 0.837 | |
| SUEBackbone=Frozen2026.05 | 89.4 | 95.3 | 96.6 | 0.161 | 0.215 | 0.225 | |
| KappaPlace-PTBackbone=Frozen2026.05 | 89.4 | 95.3 | 96.6 | 0.191 | 0.247 | 0.258 | |
| CosPlace*Backbone=ResNet101, Dim.=128, Retrained with L_lb=true2026.06 | 89.4 | 96.6 | 97.9 | — | — | — | |
| La-CosPlaceDim=+1024, Language Augmentation=true, Fusion=CAT Fusion2026.02 | 89.3 | 95 | 96.2 | — | — | — | |
| KappaPlace-JTBackbone=Fine-tuned2026.05 | 89.3 | 95.2 | 96.5 | 0.175 | 0.23 | 0.242 | |
| SPE-VLADDesc. dim.=32768, Train set=Pitts30k, Backbone=VGG-162022.04 | 89.2 | — | — | — | — | — | |
| SRFSBackbones=VGG-16, Cd=40962025.07 | 89.1 | — | — | — | — | — | |
| SFRSBackbone=VGG-16, Descriptor Dimension=40962023.08 | 89.1 | — | — | — | — | — |