Cross-view Geo-localization on DenseUAV
97.17Recall@1DINO-GFSA
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DINO-GFSABackbone=DINOv3-ViT-L2026.05 | 97.17 | 99.57 | 97.68 | |
| SHAABackbone=SwinV2-B2026.05 | 93.69 | 98.76 | 94.91 | |
| SKYPARTParams=26.95 M, Input Resolution=256×2562026.05 | 91.85 | 97.81 | 93.59 | |
| MEANVenue=TGRS’25, Params=36.5 M, Input Resolution=256×2562026.05 | 90.18 | — | — | |
| CEUSPVenue=PRCV’25, Params=∼28 M*, Input Resolution=256×2562026.05 | 89.45 | 96.05 | 91.01 | |
| MCCGBackbone=ConvNext-T2026.05 | 89.19 | 96.87 | 90.99 | |
| CAMPVenue=TGRS’24, Params=91.4 M, Input Resolution=256×2562026.05 | 88.72 | — | — | |
| Yang et al.Venue=RA-L’25, Params=∼86 M*, Input Resolution=256×2562026.05 | 86.27 | 96.83 | 88.87 | |
| DINOv2 based methodBackbone=DINOv2-B2026.05 | 86.27 | 96.83 | 88.87 | |
| MCCGVenue=TCSVT’24, Params=∼28 M*, Input Resolution=256×2562026.05 | 83.14 | 93.39 | 85.94 | |
| DenseUAV baselineVenue=TIP’23, Params=∼22 M*, Input Resolution=256×2562026.05 | 83.01 | 95.58 | 86.5 | |
| DenseUAV BaselineBackbone=ViT-S2026.05 | 83.01 | 95.58 | 86.5 | |
| FRSABackbone=ViT-S2026.05 | 81.21 | 94.55 | 85.11 | |
| LPNBackbone=ViT-S2026.05 | 71.77 | 90.13 | 77.95 | |
| Sample4GeoVenue=ICCV’23, Params=87.6 M, Input Resolution=256×2562026.05 | 49.38 | 78.29 | 61.72 | |
| MSBABackbone=ResNet502026.05 | 46.13 | 64.22 | 52.64 |