Method overview
| challenge |
pixel-level semantic labeling |
| used Cityscapes data |
fine annotations, stereo |
| used external data |
ImageNet |
| runtime |
n/a |
| subsampling |
no |
Average results
| Metric |
Value |
| IoU Classes | 81.2653 |
| iIoU Classes | 61.2169 |
| IoU Categories | 91.8522 |
| iIoU Categories | 81.0347 |
Class results
| Class |
IoU |
iIoU |
| road | 98.75 | - |
| sidewalk | 87.5812 | - |
| building | 93.9237 | - |
| wall | 60.7819 | - |
| fence | 63.5238 | - |
| pole | 70.7688 | - |
| traffic light | 78.0407 | - |
| traffic sign | 81.5587 | - |
| vegetation | 93.9712 | - |
| terrain | 73.5056 | - |
| sky | 95.803 | - |
| person | 88.0181 | 72.1697 |
| rider | 73.4848 | 54.6591 |
| car | 96.3235 | 90.6125 |
| truck | 72.6808 | 43.8061 |
| bus | 87.2303 | 56.8702 |
| train | 78.9455 | 52.8584 |
| motorcycle | 70.6944 | 52.3514 |
| bicycle | 78.4544 | 66.408 |
Category results
| Category |
IoU |
iIoU |
| flat | 98.7638 | - |
| nature | 93.6999 | - |
| object | 76.5817 | - |
| sky | 95.803 | - |
| construction | 94.197 | - |
| human | 88.0902 | 73.3307 |
| vehicle | 95.8295 | 88.7386 |
Links
Download results as .csv file
Benchmark page
