Method Details
Details for method 'HRN'
Method overview
name | HRN |
challenge | pixel-level semantic labeling |
details | Hierarchical residual network |
publication | Anonymous |
project page / code | |
used Cityscapes data | fine annotations |
used external data | |
runtime | 45 s Nvidia RTX 2080Ti |
subsampling | no |
submission date | May, 2023 |
previous submissions |
Average results
Metric | Value |
---|---|
IoU Classes | 77.6527 |
iIoU Classes | 53.4262 |
IoU Categories | 89.9834 |
iIoU Categories | 77.1841 |
Class results
Class | IoU | iIoU |
---|---|---|
road | 98.4411 | - |
sidewalk | 85.1329 | - |
building | 92.6358 | - |
wall | 55.9249 | - |
fence | 56.8465 | - |
pole | 63.8512 | - |
traffic light | 72.2387 | - |
traffic sign | 75.8889 | - |
vegetation | 93.1871 | - |
terrain | 72.2937 | - |
sky | 95.2685 | - |
person | 84.3121 | 66.2726 |
rider | 68.1211 | 44.8056 |
car | 95.474 | 89.0498 |
truck | 66.6812 | 33.3655 |
bus | 80.6608 | 44.1898 |
train | 81.6248 | 49.858 |
motorcycle | 63.299 | 42.153 |
bicycle | 73.5188 | 57.7153 |
Category results
Category | IoU | iIoU |
---|---|---|
flat | 98.6284 | - |
nature | 92.9527 | - |
object | 70.4619 | - |
sky | 95.2685 | - |
construction | 93.0163 | - |
human | 84.6712 | 67.4508 |
vehicle | 94.8847 | 86.9174 |