Method Details
Details for method 'RFLR'
Method overview
| name | RFLR |
| challenge | pixel-level semantic labeling |
| details | Random Forest with Learned Representations for Semantic Segmentation |
| publication | Random Forest with Learned Representations for Semantic Segmentation Byeongkeun Kang, Truong Q. Nguyen IEEE Transactions on Image Processing https://ieeexplore.ieee.org/document/8667358 |
| project page / code | |
| used Cityscapes data | fine annotations, coarse annotations, 16bit |
| used external data | |
| runtime | 0.03 s NVIDIA Tesla K40c |
| subsampling | 4 |
| submission date | May, 2018 |
| previous submissions |
Average results
| Metric | Value |
|---|---|
| IoU Classes | 30.0261 |
| iIoU Classes | 7.75894 |
| IoU Categories | 60.245 |
| iIoU Categories | 22.6149 |
Class results
| Class | IoU | iIoU |
|---|---|---|
| road | 82.0087 | - |
| sidewalk | 32.8849 | - |
| building | 58.8641 | - |
| wall | 9.90747 | - |
| fence | 7.10731 | - |
| pole | 10.0967 | - |
| traffic light | 11.9125 | - |
| traffic sign | 17.8771 | - |
| vegetation | 73.9562 | - |
| terrain | 43.64 | - |
| sky | 84.6223 | - |
| person | 26.6855 | 18.5004 |
| rider | 2.05551 | 0.372319 |
| car | 64.8778 | 23.5529 |
| truck | 8.85882 | 2.42782 |
| bus | 10.5796 | 4.5677 |
| train | 4.41233 | 2.02804 |
| motorcycle | 2.40554 | 0.596251 |
| bicycle | 17.743 | 10.0262 |
Category results
| Category | IoU | iIoU |
|---|---|---|
| flat | 93.7748 | - |
| nature | 73.569 | - |
| object | 15.656 | - |
| sky | 84.6223 | - |
| construction | 60.9855 | - |
| human | 29.8187 | 20.3393 |
| vehicle | 63.2887 | 24.8906 |