Object Detection on DIOR (test)
92.6mAP50DisDop-FT
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| DisDop-FTBackbone=Swin-T, Pre-Training Data=LAE-1M, Detection Type=Open-Vocabulary Object Detection2026.05 | 92.6 | — | — | — | |
| LAE-DINO-FTBackbone=Swin-T, Pre-Training Data=LAE-1M, Detection Type=Open-Vocabulary Object Detection2026.05 | 92.2 | — | — | — | |
| GroundingDINO-FTBackbone=Swin-T, Pre-Training Data=LAE-1M, Detection Type=Open-Vocabulary Object Detection2026.05 | 91.1 | — | — | — | |
| GLIP-FTBackbone=Swin-T, Pre-Training Data=LAE-1M, Detection Type=Open-Vocabulary Object Detection2026.05 | 88.9 | — | — | — | |
| MTPBackbone=Swin-H, Pre-Training Data=MillionAID, Detection Type=Generic Object Detection2026.05 | 81.1 | — | — | — | |
| SkySensePublication=CVPR’242026.04 | 78.73 | — | — | — | |
| SkySenseBackbone=Swin-H, Pre-Training Data=multi-modal RSI, Detection Type=Generic Object Detection2026.05 | 78.7 | — | — | — | |
| RemoteAgentPublication=-2026.04 | 77.8 | — | — | — | |
| RingMoBackbone=Swin-B, Pre-Training Data=RingMoPretrain, Detection Type=Generic Object Detection2026.05 | 75.9 | — | — | — | |
| SatLasBackbone=Swin-B, Pre-Training Data=SatlasPretrain, Detection Type=Generic Object Detection2026.05 | 74.1 | — | — | — | |
| Scale-MAEPublication=ICCV’232026.04 | 73.81 | — | — | — | |
| Scale-MAEBackbone=ViT-L, Pre-Training Data=FMoW, Detection Type=Generic Object Detection2026.05 | 73.8 | — | — | — | |
| GFMPublication=ICCV’232026.04 | 72.84 | — | — | — | |
| TeacherParams(MB)=30.8, BOPs(G)=21,4282022.12 | 71.95 | — | — | — | |
| TeacherParams(MB)=203, BOPs(G)=178,1762022.12 | 71.7 | — | — | — | |
| GHOSTParams(MB)=9.3, BOPs(G)=2,1582022.12 | 71.53 | — | — | — | |
| TOVBackbone=ResNet-50, Pre-Training Data=TOV-NI,TOV-R, Detection Type=Generic Object Detection2026.05 | 70.2 | — | — | — | |
| ARSDParams(MB)=52, BOPs(G)=42,5982022.12 | 70.1 | — | — | — | |
| FMSSDParams(MB)=5442022.12 | 69.5 | — | — | — | |
| O2-DNetParams(MB)=8362022.12 | 68.3 | — | — | — | |
| GFLParams(MB)=76, BOPs(G)=99,7682022.12 | 68 | — | — | — | |
| ATSSParams(MB)=75, BOPs(G)=97,7922022.12 | 67.7 | — | — | — | |
| FCOSParams(MB)=127, BOPs(G)=126,4742022.12 | 67.6 | — | — | — | |
| GASSLBackbone=ResNet-50, Detection Type=Generic Object Detection2026.05 | 67.4 | — | — | — | |
| CACOBackbone=ResNet-50, Pre-Training Data=Sentinel-2, Detection Type=Generic Object Detection2026.05 | 66.9 | — | — | — | |
| RetainNetParams(MB)=221, BOPs(G)=184,9542022.12 | 65.7 | — | — | — | |
| ShuffleNetParams(MB)=48, BOPs(G)=89,4052022.12 | 61.3 | — | — | — | |
| MobileNetV2Params(MB)=41, BOPs(G)=77,9262022.12 | 58.2 | — | — | — | |
| YOLOv3Params(MB)=247, BOPs(G)=125,1532022.12 | 57.1 | — | — | — | |
| FalconPublication=arXiv’252026.04 | 56.65 | — | — | — | |
| Faster R-CNNParams(MB)=240, BOPs(G)=186,5722022.12 | 54.1 | — | — | — | |
| FLAME with RS-OWL-ViT-v2Evaluation protocol=Few-shot2025.10 | 53.21 | — | — | — | |
| SIoUEvaluation protocol=Few-shot2025.10 | 52.85 | — | — | — | |
| RS-OWL-ViT-v2Evaluation protocol=Zero-shot2025.10 | 29.39 | — | — | — | |
| Florence-2-LPublication=CVPR’242026.04 | 26.98 | — | — | — | |
| OWL-ViT-v2Evaluation protocol=Zero-shot2025.10 | 14.98 | — | — | — | |
| FG-CLIPOpen-vocabulary=true, Framework=CastDet2026.03 | — | 10.9 | 64.9 | 54.1 | |
| GeoAlignCLIPOpen-vocabulary=true, Framework=CastDet2026.03 | — | 17.1 | 65 | 55.4 | |
| GeoRSCLIPOpen-vocabulary=true, Framework=CastDet2026.03 | — | 14 | 64.4 | 54.3 | |
| GLIPOpen-vocabulary=true, Framework=CastDet2026.03 | — | 5.5 | 47.9 | 39.4 | |
| GroundingDINOOpen-vocabulary=true, Framework=CastDet2026.03 | — | 9.8 | 48.7 | 39.5 | |
| LRSCLIPOpen-vocabulary=true, Framework=CastDet2026.03 | — | 15.2 | 64.5 | 54.7 | |
| ViLDOpen-vocabulary=true, Framework=CastDet2026.03 | — | 0.7 | 47.7 | 39.3 |