Method Details
Details for method 'Learnable Tree Filter'
Method overview
| name | Learnable Tree Filter |
| challenge | pixel-level semantic labeling |
| details | Learnable Tree Filter for Structure-preserving Feature Transform |
| publication | Learnable Tree Filter for Structure-preserving Feature Transform Lin Song; Yanwei Li; Zeming Li; Gang Yu; Hongbin Sun; Jian Sun; Nanning Zheng NeurIPS 2019 http://papers.neurips.cc/paper/8448-learnable-tree-filter-for-structure-preserving-feature-transform |
| project page / code | https://github.com/StevenGrove/TreeFilter-Torch |
| used Cityscapes data | fine annotations |
| used external data | ImageNet |
| runtime | n/a |
| subsampling | no |
| submission date | May, 2019 |
| previous submissions |
Average results
| Metric | Value |
|---|---|
| IoU Classes | 80.7736 |
| iIoU Classes | 60.7438 |
| IoU Categories | 91.5841 |
| iIoU Categories | 81.0932 |
Class results
| Class | IoU | iIoU |
|---|---|---|
| road | 98.6991 | - |
| sidewalk | 86.839 | - |
| building | 93.41 | - |
| wall | 53.5526 | - |
| fence | 60.2406 | - |
| pole | 69.5919 | - |
| traffic light | 77.1022 | - |
| traffic sign | 81.09 | - |
| vegetation | 93.6783 | - |
| terrain | 71.714 | - |
| sky | 95.8196 | - |
| person | 87.5253 | 71.8796 |
| rider | 72.1534 | 51.103 |
| car | 96.1772 | 91.0823 |
| truck | 72.668 | 44.1579 |
| bus | 87.0108 | 55.4378 |
| train | 86.3689 | 56.4093 |
| motorcycle | 72.6169 | 51.5156 |
| bicycle | 78.4403 | 64.3648 |
Category results
| Category | IoU | iIoU |
|---|---|---|
| flat | 98.7449 | - |
| nature | 93.3766 | - |
| object | 75.6508 | - |
| sky | 95.8196 | - |
| construction | 93.7924 | - |
| human | 88.0053 | 72.9049 |
| vehicle | 95.6992 | 89.2816 |