Semantic Segmentation on BDD to UAVID (test)
35.4RoadEAGLE
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| EAGLESetting=Open-Vocab Seg2024.06 | 35.4 | 68.9 | 50.6 | 59.2 | 51.7 | 38.6 | 50.7 | |
| FreSeg + Cross-ViewSetting=Open-Vocab Seg2024.06 | 32.6 | 67.3 | 47.9 | 51.8 | 50.3 | 37.2 | 47.9 | |
| EAGLESetting=Unsupervised Domain Adaptation, Backbone=DAFormer2024.06 | 29 | 66.1 | 41.5 | 55.6 | 53.3 | 21.5 | 44.5 | |
| DenseCLIP + Cross-ViewSetting=Open-Vocab Seg2024.06 | 25.9 | 60.9 | 39.5 | 35.5 | 47.1 | 33.9 | 40.5 | |
| DAFormerSetting=Unsupervised Domain Adaptation2024.06 | 25.8 | 65.4 | 38.7 | 54.5 | 51.3 | 14.8 | 41.8 | |
| EAGLESetting=Unsupervised Domain Adaptation, Backbone=DeepLab2024.06 | 24 | 53.8 | 39 | 52.2 | 48.3 | 16.9 | 39 | |
| Polar Trans.Setting=Unsupervised Domain Adaptation2024.06 | 21.1 | 9.6 | 36.4 | 24.1 | 14.6 | 4.6 | 18.4 | |
| BiMaLSetting=Unsupervised Domain Adaptation2024.06 | 19.5 | 52.4 | 35.1 | 50.4 | 46 | 10.2 | 35.6 | |
| No AdaptationSetting=No Adaptation2024.06 | 19.2 | 8.5 | 34.6 | 18.4 | 13.6 | 4 | 16.4 |