Method Details
Details for method 'DF-Seg'
Method overview
| name | DF-Seg |
| challenge | pixel-level semantic labeling |
| details | DF1-Seg-d8 |
| publication | Partial Order Pruning: for Best Speed/Accuracy Trade-off in Neural Architecture Search Xin Li, Yiming Zhou, Zheng Pan, Jiashi Feng CVPR 2019 https://arxiv.org/abs/1903.03777 |
| project page / code | https://github.com/lixincn2015/Partial-Order-Pruning |
| used Cityscapes data | fine annotations |
| used external data | ImageNet |
| runtime | 0.007 s GTX 1080Ti |
| subsampling | no |
| submission date | November, 2018 |
| previous submissions |
Average results
| Metric | Value |
|---|---|
| IoU Classes | 71.3865 |
| iIoU Classes | 45.0028 |
| IoU Categories | 86.647 |
| iIoU Categories | 69.6295 |
Class results
| Class | IoU | iIoU |
|---|---|---|
| road | 97.9251 | - |
| sidewalk | 81.5939 | - |
| building | 90.5711 | - |
| wall | 44.829 | - |
| fence | 50.429 | - |
| pole | 53.2177 | - |
| traffic light | 62.6664 | - |
| traffic sign | 67.9814 | - |
| vegetation | 91.6503 | - |
| terrain | 68.3179 | - |
| sky | 93.9324 | - |
| person | 78.7614 | 55.9361 |
| rider | 58.9193 | 34.6702 |
| car | 93.4132 | 84.3142 |
| truck | 50.5486 | 26.1686 |
| bus | 71.4029 | 39.3085 |
| train | 77.4415 | 37.8834 |
| motorcycle | 56.3738 | 29.936 |
| bicycle | 66.3694 | 51.8052 |
Category results
| Category | IoU | iIoU |
|---|---|---|
| flat | 98.1307 | - |
| nature | 91.2611 | - |
| object | 61.0083 | - |
| sky | 93.9324 | - |
| construction | 90.9803 | - |
| human | 78.8888 | 57.0874 |
| vehicle | 92.3275 | 82.1717 |