Visual Place Recognition on Oxford
98.4Recall@1DINOv2-RIA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DINOv2-RIALearning Protocol=training-free (with pre-trained backbone)2026.01 | 98.4 | 100 | |
| AnyLoc-VLAD-DINOv2-PCALearning Protocol=Unsupervised Learning (with pre-trained backbone)2026.01 | 96.9 | 100 | |
| CosPlaceLearning Protocol=Supervised Learning2026.01 | 95.3 | 99.5 | |
| MixVPRLearning Protocol=Supervised Learning2026.01 | 92.7 | 99.5 | |
| DINOv2-GeMLearning Protocol=training-free (with pre-trained backbone)2026.01 | 88.4 | 94.7 | |
| DINOv2-VLADLearning Protocol=Unsupervised Learning (with pre-trained backbone)2026.01 | 82.7 | 92.6 | |
| AnyLoc-VLAD-DINO-PCALearning Protocol=Unsupervised Learning (with pre-trained backbone)2026.01 | 82.2 | 99 | |
| NetVLADLearning Protocol=Supervised Learning2026.01 | 57.6 | 79.1 |