Text-to-Satellite Image Retrieval on CVG-Text Tokyo
75.92R@1MAPS
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MAPSQuery=Joint query2026.06 | 75.92 | 79.17 | |
| Pairwise alignmentQuery=Joint query2026.06 | 71.67 | 75.33 | |
| GRAMQuery=Joint query2026.06 | 70.24 | 73.68 | |
| PMRLQuery=Joint query2026.06 | 69.81 | 74.02 | |
| VASTQuery=Joint query2026.06 | 68.92 | 72.31 | |
| CLIP-L/14Query=Image only2026.06 | 67.92 | 70.67 | |
| BLIPQuery=Image only2026.06 | 66.25 | 69.08 | |
| CLIP-B/16Query=Image only2026.06 | 62.33 | 65.25 | |
| PLANETQuery=Text only2026.06 | 43.42 | 46.42 | |
| CrossText2LocQuery=Text only2026.06 | 41.75 | 43.86 | |
| OursInternVL# Trainable Params=9M2026.04 | 39.17 | 44.33 | |
| CrossText2Loc# Trainable Params=428M2026.04 | 36.83 | 39.58 | |
| CrossText2Loc# Trainable Params=428M2026.04 | 34.33 | 38.33 | |
| OursInternVL# Trainable Params=9M2026.04 | 33 | 36.5 | |
| BLIP# Trainable Params=470M2026.04 | 30.67 | 34.5 | |
| BLIP# Trainable Params=470M2026.04 | 29.75 | 33.67 | |
| BLIPQuery=Text only2026.06 | 29.75 | 33.67 | |
| CLIP-L/14# Trainable Params=428M2026.04 | 28.08 | 30.5 | |
| CLIP-L/14Query=Text only2026.06 | 28.08 | 30.5 | |
| CLIP-B/16Query=Text only2026.06 | 24 | 27.25 | |
| CLIP-L/14# Trainable Params=428M2026.04 | 21 | 23.17 | |
| SigLIP-B/16# Trainable Params=203M2026.04 | 15.58 | 17.92 | |
| X-VLM# Trainable Params=216M2026.04 | 12.46 | 14.34 | |
| SigLIP-B/16# Trainable Params=203M2026.04 | 11.92 | 13.67 | |
| ViLT# Trainable Params=87.4M2026.04 | 10.83 | 15.5 | |
| X-VLM# Trainable Params=216M2026.04 | 9.53 | 10.94 | |
| ViLT# Trainable Params=87.4M2026.04 | 4.67 | 9.17 |