Method Details
Details for method 'HRN+DCNv2_for_DOAS'
Method overview
| name | HRN+DCNv2_for_DOAS |
| challenge | pixel-level semantic labeling |
| details | HRN with DCNv2 for DOAS in paper "Dynamic Obstacle Avoidance System based on Rapid Instance Segmentation Network" |
| publication | Anonymous |
| project page / code | |
| used Cityscapes data | fine annotations |
| used external data | COCO2017 |
| runtime | 0.032 s Nvidia RTX 2080Ti |
| subsampling | no |
| submission date | May, 2023 |
| previous submissions |
Average results
| Metric | Value |
|---|---|
| IoU Classes | 81.176 |
| iIoU Classes | 59.7751 |
| IoU Categories | 91.6349 |
| iIoU Categories | 80.9724 |
Class results
| Class | IoU | iIoU |
|---|---|---|
| road | 98.6285 | - |
| sidewalk | 86.6908 | - |
| building | 93.6834 | - |
| wall | 62.0716 | - |
| fence | 63.9203 | - |
| pole | 70.4162 | - |
| traffic light | 76.9773 | - |
| traffic sign | 79.4727 | - |
| vegetation | 93.7854 | - |
| terrain | 72.8914 | - |
| sky | 95.5622 | - |
| person | 87.6913 | 72.0263 |
| rider | 73.3741 | 51.9255 |
| car | 96.03 | 90.2719 |
| truck | 71.169 | 40.3289 |
| bus | 87.6827 | 52.8925 |
| train | 82.9019 | 55.6602 |
| motorcycle | 71.4285 | 48.5567 |
| bicycle | 77.9665 | 66.5387 |
Category results
| Category | IoU | iIoU |
|---|---|---|
| flat | 98.7498 | - |
| nature | 93.591 | - |
| object | 75.9304 | - |
| sky | 95.5622 | - |
| construction | 93.9798 | - |
| human | 87.8835 | 73.2706 |
| vehicle | 95.7476 | 88.6742 |