Method Details
Details for method 'NfS-Seg'
Method overview
| name | NfS-Seg |
| challenge | pixel-level semantic labeling |
| details | |
| publication | Uncertainty-Aware Knowledge Distillation for Real-Time Scene Segmentation: 7.43 GFLOPs at Full-HD Image with 120 fps |
| project page / code | |
| used Cityscapes data | fine annotations, coarse annotations, stereo, video |
| used external data | ImageNet |
| runtime | 0.00837312 s NPU |
| subsampling | no |
| submission date | October, 2018 |
| previous submissions |
Average results
| Metric | Value |
|---|---|
| IoU Classes | 73.0734 |
| iIoU Classes | 44.4281 |
| IoU Categories | 87.5059 |
| iIoU Categories | 70.0605 |
Class results
| Class | IoU | iIoU |
|---|---|---|
| road | 98.2466 | - |
| sidewalk | 83.7212 | - |
| building | 91.2922 | - |
| wall | 55.5916 | - |
| fence | 53.0796 | - |
| pole | 57.2074 | - |
| traffic light | 62.5839 | - |
| traffic sign | 68.7317 | - |
| vegetation | 92.3823 | - |
| terrain | 70.8857 | - |
| sky | 94.5671 | - |
| person | 78.87 | 54.4235 |
| rider | 59.7102 | 35.7349 |
| car | 93.9508 | 86.2932 |
| truck | 60.6977 | 26.2324 |
| bus | 76.3189 | 37.8588 |
| train | 67.302 | 36.4646 |
| motorcycle | 55.4358 | 29.9029 |
| bicycle | 67.8191 | 48.5142 |
Category results
| Category | IoU | iIoU |
|---|---|---|
| flat | 98.4154 | - |
| nature | 92.0371 | - |
| object | 63.9645 | - |
| sky | 94.5671 | - |
| construction | 91.3918 | - |
| human | 79.0141 | 56.1444 |
| vehicle | 93.1513 | 83.9766 |