Semantic Segmentation on Cross-City Tokyo
83.4Road IoUCross-City
Evaluation results
| Method | Links | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Cross-City2018.02 | 83.4 | 35.4 | 72.8 | 12.3 | 12.7 | 77.4 | 64.3 | 42.7 | 21.5 | 64.1 | 20.8 | 8.9 | 40.3 | 42.8 | |
| BaselineBackbone=ResNet-101, Adaptation Level=None2018.02 | 82.9 | 31.3 | 78.7 | 14.2 | 24.5 | 81.6 | 89.2 | 48.6 | 33.3 | 70.5 | 7.7 | 11.5 | 45.9 | 47.7 | |
| AdaptSegNetBackbone=ResNet-101, Adaptation Level=feature space2018.02 | 81.5 | 30.8 | 76.6 | 15.3 | 20.2 | 82 | 84 | 49.4 | 33.3 | 70.5 | 4.5 | 24.3 | 51.6 | 48 | |
| AdaptSegNetBackbone=ResNet-101, Adaptation Level=output space2018.02 | 81.5 | 26 | 77.8 | 17.8 | 26.8 | 82.7 | 90.9 | 55.8 | 38 | 72.1 | 4.2 | 24.5 | 50.8 | 49.9 |