Ship Re-identification on HOSS-ReID (All)
60.5mAPDRI-L
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| DRI-LBackbone=DINOv3, Trainable Params (M)=7.052025.12 | 60.5 | 69.9 | 79 | 85.2 | |
| DRI-SBackbone=DINOv3, Trainable Params (M)=1.542025.12 | 57.9 | 67 | 80.1 | 87.5 | |
| TransOSSBackbone=Transformer, Trainable Params (M)=86.00, Pre-training=With optical-SAR image pairs2025.12 | 57.4 | 65.9 | 79.5 | 85.8 | |
| DRI-BBackbone=DINOv3, Trainable Params (M)=2.932025.12 | 56.1 | 65.9 | 76.7 | 83 | |
| TransOSSBackbone=Transformer, Trainable Params (M)=86.00, Pre-training=Without optical-SAR image pairs2025.12 | 49.4 | 61.9 | 71 | 78.4 | |
| VersReIDBackbone=Transformer, Trainable Params (M)=85.902025.12 | 49.3 | 59.7 | 70.5 | 78.4 | |
| TransReIDBackbone=Transformer, Trainable Params (M)=86.002025.12 | 48.1 | 60.8 | 69.3 | 73.9 | |
| DeiT-baseBackbone=Transformer, Trainable Params (M)=86.002025.12 | 47.2 | 58.1 | 69.6 | 74.1 | |
| MCJABackbone=ResNet, Trainable Params (M)=26.402025.12 | 47.1 | 59.1 | 67.9 | 73 | |
| DEENBackbone=ResNet2025.12 | 43.8 | 58.5 | 64.2 | 66.5 | |
| AGWBackbone=ResNet, Trainable Params (M)=33.722025.12 | 43.6 | 57.4 | 64.2 | 68.8 | |
| ViT-baseBackbone=Transformer, Trainable Params (M)=86.002025.12 | 43 | 56.2 | 64.8 | 69.9 | |
| SOLIDERBackbone=Transformer2025.12 | 38.2 | 50.6 | 63.1 | 69.9 | |
| ViT-SBackbone=DINOv3, Trainable Params (M)=0.142025.12 | 34.4 | 52.3 | 58.5 | 60.2 | |
| Hc-TriBackbone=ResNet2025.12 | 34 | 47.2 | 54.6 | 59.7 | |
| ViT-BBackbone=DINOv3, Trainable Params (M)=0.282025.12 | 33.1 | 48.3 | 56.8 | 60.8 | |
| LbABackbone=ResNet, Trainable Params (M)=23.522025.12 | 33 | 48.3 | 59.7 | 62.5 | |
| CM-NASBackbone=ResNet2025.12 | 30.7 | 46 | 54.6 | 57.4 | |
| ViT-LBackbone=DINOv3, Trainable Params (M)=0.372025.12 | 28.9 | 42 | 50 | 53.4 |