Semantic Segmentation on Cityscapes 1.0 (val) (Robustness Metrics)
83.6Clean ScoreRSSeg-CSWin-S
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| RSSeg-CSWin-SBackbone=CSWin-S2022.12 | 83.6 | 66.8 | 79.9 | 69.5 | 59 | 75.3 | 63.4 | |
| SegFormer-B5Backbone=MiT-B52022.12 | 82.4 | 65.8 | 79.9 | 67.9 | 61.6 | 74.4 | 59.4 | |
| FAN-L-HybridBackbone=Hybrid2022.12 | 82.3 | 68.7 | 83.5 | — | — | — | — | |
| CSWin-SBackbone=CSWin-S2022.12 | 81.9 | 66.4 | 81.1 | 68.7 | 61.5 | 74.2 | 61.2 | |
| RSSeg-ViT-LBackbone=ViT-L2022.12 | 81.3 | 70.1 | 86.3 | 72.1 | 70.3 | 75.7 | 62.5 | |
| RSSeg-ViT-BBackbone=ViT-B2022.12 | 81.2 | 67.9 | 83.6 | 70.2 | 65.4 | 74.4 | 61.4 | |
| GSCNN+Backbone=Not specified2022.12 | 80.9 | 40 | 49.4 | 54.8 | 9.9 | 55.1 | 40.1 | |
| DLV3+Backbone=ResNet-1012022.12 | 80.9 | 43.9 | 54.2 | 53.4 | 22.8 | 58.3 | 41.1 | |
| ViT-Linear-BBackbone=ViT-B2022.12 | 79.5 | 65.9 | 82.9 | 69.1 | 63.2 | 71.8 | 59.6 | |
| SETR-PUP-LBackbone=ViT-L2022.12 | 79.3 | 66.1 | 83.3 | 70.5 | 63.4 | 70.3 | 60.1 |