Image-text retrieval on UCM
20.48I2T R@1RemoteCLIP (ViT-B)
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| RemoteCLIP (ViT-B)Vision Backbone=ViT-B, Finetuning Data=RET-3 + DET-10 + SEG-4, Total Params (M)=1512025.12 | 20.48 | 59.85 | 83.33 | 18.67 | 61.52 | 94.29 | 56.36 | |
| GeoMELTVision Backbone=MWT, Finetuning Data=Multi-Task Data, Total Params (M)=2712025.12 | 20.48 | 57.14 | 84.76 | 18.86 | 64.29 | 94.95 | 56.75 | |
| RemoteCLIP (ViT-L)Vision Backbone=ViT-L, Finetuning Data=RET-3 + DET-10 + SEG-4, Total Params (M)=4282025.12 | 19.05 | 54.29 | 80.95 | 17.71 | 62.19 | 93.9 | 54.68 | |
| Rahhal et al.Vision Backbone=ViT-B, Finetuning Data=UCM, Total Params (M)=1512025.12 | 19.04 | 53.33 | 77.61 | 19.33 | 64 | 91.42 | 54.12 | |
| AMFMNVision Backbone=ResNet50, Finetuning Data=UCM, Total Params (M)=25*2025.12 | 16.67 | 45.71 | 68.57 | 12.86 | 53.24 | 79.43 | 46.08 | |
| CLIP-RSICDVision Backbone=ViT-B, Finetuning Data=RSICD, Total Params (M)=1512025.12 | 15.71 | 50 | 82.38 | 13.81 | 57.05 | 91.24 | 51.7 | |
| CLIP (ViT-L)Vision Backbone=ViT-L, Finetuning Data=Zero-shot, Total Params (M)=4282025.12 | 11.91 | 43.33 | 65.24 | 10.76 | 46 | 73.33 | 41.76 | |
| KCRVision Backbone=ResNet18, Finetuning Data=UCM, Total Params (M)=11*2025.12 | 11.9 | 48.57 | 71.43 | 17.24 | 56.95 | 81.14 | 47.87 | |
| CLIP (ViT-B)Vision Backbone=ViT-B, Finetuning Data=Zero-shot, Total Params (M)=1512025.12 | 8.1 | 34.29 | 53.33 | 8.67 | 36.48 | 60.57 | 33.57 |