Visual Place Recognition on 17 Places
65Recall@1AnyLoc-VLAD-DINOv2
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| AnyLoc-VLAD-DINOv2Aggregation=VLAD2023.08 | 65 | 80.5 | |
| AnyLoc-VLAD-DINOv2Aggregation=VLAD, PCA-Whitening=true2023.08 | 64.8 | 81 | |
| AnyLoc-VLAD-DINOv2-PCALearning Protocol=Unsupervised Learning (with pre-trained backbone)2026.01 | 64.8 | 81 | |
| DINOv2-RIALearning Protocol=training-free (with pre-trained backbone)2026.01 | 64.7 | 81.2 | |
| DINOv2-VLADLearning Protocol=Unsupervised Learning (with pre-trained backbone)2026.01 | 64.2 | 81.2 | |
| MixVPR2023.08 | 63.8 | 78.8 | |
| AnyLoc-VLAD-DINOAggregation=VLAD2023.08 | 63.8 | 78.8 | |
| MixVPRLearning Protocol=Supervised Learning2026.01 | 63.8 | 78.8 | |
| AnyLoc-VLAD-DINO-PCALearning Protocol=Unsupervised Learning (with pre-trained backbone)2026.01 | 63.8 | 78.8 | |
| DINOv2-GeMLearning Protocol=training-free (with pre-trained backbone)2026.01 | 63.7 | 80 | |
| AnyLoc-GeM-DINOv2Aggregation=GeM pooling2023.08 | 63.6 | 79.6 | |
| AnyLoc-VLAD-DINOAggregation=VLAD, PCA-Whitening=true2023.08 | 63.3 | 78.8 | |
| DINO-CLSAggregation=CLS token2023.08 | 61.8 | 76.4 | |
| DINOv2-CLSAggregation=CLS token2023.08 | 61.8 | 78.8 | |
| NetVLAD2023.08 | 61.6 | 77.8 | |
| NetVLADLearning Protocol=Supervised Learning2026.01 | 61.6 | 77.8 | |
| CosPlace2023.08 | 61.1 | 76.1 | |
| CosPlaceLearning Protocol=Supervised Learning2026.01 | 61.1 | 76.1 | |
| CLIP-CLSAggregation=CLS token2023.08 | 59.4 | 77.6 |