Scene Parsing on Cityscapes fine-annotation (test)
80.3mIoU (Class)PAD-Net-ResNet101
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| PAD-Net-ResNet101Backbone=ResNet-101, Training Data=fine-annotation2018.05 | 80.3 | 58.8 | 90.8 | 78.5 | |
| PSPNetTraining Data=fine-annotation2018.05 | 78.4 | 56.7 | 90.6 | 78.6 | |
| PiecewiseTraining Data=fine-annotation2018.05 | 71.6 | 51.7 | 87.3 | 74.1 | |
| DeepLabTraining Data=fine-annotation2018.05 | 70.4 | 42.6 | 86.4 | 67.7 | |
| LRRTraining Data=fine-annotation2018.05 | 69.7 | 48 | 88.2 | 74.7 | |
| Dilation10Training Data=fine-annotation2018.05 | 67.1 | 42 | 86.5 | 71.1 | |
| DPNTraining Data=fine-annotation2018.05 | 66.8 | 39.1 | 86 | 69.1 | |
| SiCNNTraining Data=fine-annotation2018.05 | 66.3 | 44.9 | 85 | 71.2 | |
| CRF-RNNTraining Data=fine-annotation2018.05 | 62.5 | 34.4 | 82.7 | 66 | |
| SegNetTraining Data=fine-annotation2018.05 | 56.1 | 34.2 | 79.8 | 65.4 |