Method Details
Details for method 'ESANet RGB'
Method overview
| name | ESANet RGB |
| challenge | pixel-level semantic labeling |
| details | ESANet: Efficient RGB-D Semantic Segmentation for Indoor Scene Analysis. ESANet-R34-NBt1D using RGB images only. |
| publication | Efficient RGB-D Semantic Segmentation for Indoor Scene Analysis Daniel Seichter, Mona Köhler, Benjamin Lewandowski, Tim Wengefeld and Horst-Michael Gross |
| project page / code | https://github.com/TUI-NICR/ESANet |
| used Cityscapes data | fine annotations |
| used external data | ImageNet |
| runtime | 0.1205 s NVIDIA Jetson AGX Xavier (Jetpack 4.4, TensorRT 7.1, Float16) |
| subsampling | no |
| submission date | November, 2020 |
| previous submissions |
Average results
| Metric | Value |
|---|---|
| IoU Classes | 77.5574 |
| iIoU Classes | 53.1298 |
| IoU Categories | 90.1509 |
| iIoU Categories | 76.1974 |
Class results
| Class | IoU | iIoU |
|---|---|---|
| road | 98.4413 | - |
| sidewalk | 84.9052 | - |
| building | 92.713 | - |
| wall | 55.3943 | - |
| fence | 58.9471 | - |
| pole | 64.6633 | - |
| traffic light | 71.7233 | - |
| traffic sign | 75.7924 | - |
| vegetation | 93.3104 | - |
| terrain | 70.956 | - |
| sky | 95.2586 | - |
| person | 84.8994 | 64.1194 |
| rider | 67.6772 | 43.5926 |
| car | 95.7449 | 89.0103 |
| truck | 64.6938 | 34.1028 |
| bus | 79.1171 | 46.6804 |
| train | 80.9395 | 45.8106 |
| motorcycle | 64.4847 | 42.9047 |
| bicycle | 73.9294 | 58.8176 |
Category results
| Category | IoU | iIoU |
|---|---|---|
| flat | 98.6452 | - |
| nature | 92.9936 | - |
| object | 70.9705 | - |
| sky | 95.2586 | - |
| construction | 93.0572 | - |
| human | 85.1016 | 65.4723 |
| vehicle | 95.0295 | 86.9225 |