Visual Place Recognition On 17 Places
평가 지표
Recall@1
평가 결과
이 벤치마크에서 각 모델의 성능 결과
모델 이름 | Recall@1 | Paper Title | Repository |
|---|---|---|---|
| NetVLAD | 61.58 | NetVLAD: CNN architecture for weakly supervised place recognition | |
| AnyLoc-VLAD-DINOv2 | 65.02 | AnyLoc: Towards Universal Visual Place Recognition | |
| DINO | 61.82 | Emerging Properties in Self-Supervised Vision Transformers | |
| CLIP | 59.36 | AnyLoc: Towards Universal Visual Place Recognition | |
| CosPlace | 61.08 | Rethinking Visual Geo-localization for Large-Scale Applications | |
| SegVLAD-FineT (M) | 95.3 | Revisit Anything: Visual Place Recognition via Image Segment Retrieval | |
| MixVPR | 63.79 | MixVPR: Feature Mixing for Visual Place Recognition | |
| DINOv2 | 61.82 | DINOv2: Learning Robust Visual Features without Supervision |
0 of 8 row(s) selected.