Semantic Segmentation on Cityscapes 2012 (val) (Resolution-Scaled mIoU)
70.33mIoU (1/8 Scale)Ours (ELN)
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Ours (ELN)SegNet=DeepLab v3+, Backbone=ResNet-502022.04 | 70.33 | 73.52 | 75.33 | — | |
| Alonso et al.SegNet=DeepLab v3+, Backbone=ResNet-502022.04 | 70 | 71.6 | — | 74.2 | |
| Xin et al.SegNet=DeepLab v3+, Backbone=ResNet-502022.04 | 69.7 | 72.7 | — | 77.5 | |
| ECSSegNet=DeepLab v3+, Backbone=ResNet-502022.04 | 67.38 | 70.7 | 72.89 | 74.76 | |
| Alonso et al.SegNet=DeepLab v2, Backbone=ResNet-1012022.04 | 63 | 64.8 | — | 66.4 | |
| CutMixSegNet=DeepLab v2, Backbone=ResNet-1012022.04 | 60.34 | 63.87 | — | 67.68 | |
| BaselineSegNet=DeepLab v3+, Backbone=ResNet-502022.04 | 59.88 | 61.86 | 67.63 | 77.7 | |
| S4GANSegNet=DeepLab v2, Backbone=ResNet-1012022.04 | 59.3 | 61.9 | — | 65.8 |