Object Detection on HOD-1
93.6mAP@.5SDCM
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| SDCMBackbone=DarkNet532025.12 | 93.6 | 93 | 94.8 | 83.5 | 97.5 | 97.6 | 94.9 | 89 | 98.4 | |
| HOD-1Backbone=VGG16, Transfer Learning=true2025.12 | 89.3 | — | — | — | — | — | — | — | — | |
| Libra RCNNBackbone=ResNet502025.12 | 87.1 | 90.3 | 91.8 | 62.4 | 96.9 | 95.3 | 79.3 | 85.5 | 95.5 | |
| S2ADetBackbone=DarkNet532025.12 | 86.6 | 83.1 | 88.3 | 51.5 | 98.4 | 98 | 91.2 | 83.3 | 98.7 | |
| MethaneMapperBackbone=ResNet502025.12 | 84.5 | 82.1 | 90.4 | 60.9 | 89.4 | 87.5 | 90.6 | 86.1 | 89.1 | |
| Cascade RCNNBackbone=ResNet502025.12 | 84.3 | 89.6 | 78.6 | 55.1 | 94.9 | 94 | 88.3 | 80.8 | 93.3 | |
| HOD-1Backbone=VGG162025.12 | 83.5 | — | — | — | — | — | — | — | — | |
| Double-Head RCNNBackbone=ResNet502025.12 | 81.2 | — | — | — | — | — | — | — | — | |
| FCOSBackbone=ResNet502025.12 | 80.9 | — | — | — | — | — | — | — | — | |
| YOLOv5Backbone=DarkNet532025.12 | 80.4 | 82.4 | 80.3 | 34 | 97.7 | 97.6 | 76.9 | 81.1 | 93.3 | |
| YOLOFBackbone=DarkNet532025.12 | 80.3 | 87.4 | 75 | 39.7 | 96.4 | 90.6 | 78.2 | 82.3 | 92.6 | |
| FoveaBoxBackbone=ResNet502025.12 | 80.2 | — | — | — | — | — | — | — | — | |
| Faster RCNNBackbone=ResNet502025.12 | 78.8 | 72.2 | 80 | 44.7 | 95.1 | 92.4 | 80.2 | 72.8 | 93.1 | |
| Deformable DETRBackbone=VGG162025.12 | 71.7 | 71.4 | 70.4 | 56.2 | 87.4 | 80.6 | 83.8 | 46.8 | 76.6 | |
| RetinanetBackbone=VGG162025.12 | 51.6 | 42.4 | 58.9 | 24.8 | 43.4 | 83.3 | 49.1 | 31 | 80 |