Semantic Segmentation on Cityscapes 12 (val test)
28.9mIoUCurriculum Domain Adaptation (I+SP)
Evaluation results
| Method | Links | ||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Curriculum Domain Adaptation (I+SP)source_dataset=GTA, curriculum_level=Image + Superpixel2017.07 | 28.9 | 14.6 | 11.9 | 6 | 11.1 | 8.4 | 16.8 | 16.3 | 66.5 | 18.9 | 9.3 | 75.7 | 13.3 | 0 | 71.7 | 55.2 | 38 | 18.8 | 22 | 74.9 | |
| Curriculum Domain Adaptation (SP)source_dataset=GTA, curriculum_level=Superpixel-level2017.07 | 27.8 | 15.6 | 11.7 | 5.7 | 12 | 9.2 | 12.9 | 15.5 | 64.9 | 15.5 | 9.1 | 74.6 | 11.1 | 0 | 70.5 | 56.1 | 34.8 | 15.9 | 21.8 | 72.1 | |
| FCN Wldsource_dataset=GTA, adaptation=Global, reference=[27]2017.07 | 27.1 | 0 | 5.4 | 14.9 | 2.7 | 10.9 | 3.5 | 14.2 | 64.6 | 7.3 | 4.2 | 79.2 | 21.3 | 0 | 62.1 | 70.4 | 44.1 | 8 | 32.4 | 70.4 | |
| SPsource_dataset=GTA, component=Superpixel curriculum2017.07 | 26.8 | 0.3 | 4.1 | 7.6 | 0 | 0.2 | 0.9 | 0 | 81.6 | 25.3 | 3.5 | 73 | 32.1 | 0 | 71 | 61.9 | 26.2 | 30.4 | 19.2 | 71.8 | |
| Curriculum Domain Adaptation (I)source_dataset=GTA, curriculum_level=Image-level2017.07 | 23.1 | 9.5 | 9.4 | 10.2 | 14 | 20.2 | 3.8 | 13.6 | 63.8 | 3.4 | 10.6 | 56.9 | 2.8 | 10.9 | 69.7 | 60.5 | 31.8 | 10.9 | 10.8 | 26.4 | |
| NoAdaptsource_dataset=GTA, adaptation=None, implementation=Local2017.07 | 22.3 | 13.8 | 8.7 | 7.3 | 16.8 | 21 | 4.3 | 14.9 | 64.4 | 5 | 17.5 | 45.9 | 2.4 | 6.9 | 64.1 | 55.3 | 41.6 | 8.4 | 6.8 | 18.1 | |
| SP Lndmksource_dataset=GTA, component=Superpixel Landmarks2017.07 | 21.4 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 82.9 | 10 | 0 | 74.5 | 22.5 | 0 | 69.9 | 52.7 | 13.1 | 11.2 | 8 | 61.8 | |
| NoAdaptsource_dataset=GTA, adaptation=None, reference=[27]2017.07 | 21.1 | 0 | 3.1 | 7.4 | 1 | 16 | 0 | 10.4 | 58.9 | 3.7 | 1 | 76.5 | 13 | 0 | 47.7 | 67.1 | 36 | 9.5 | 18.9 | 31.9 |