Method Details
Details for method 'SDBNetV2'
Method overview
| name | SDBNetV2 |
| challenge | pixel-level semantic labeling |
| details | |
| publication | Improved Short-term Dense Bottleneck network for efficient scene analysis Tanmay Singha; Duc-Son Pham; Aneesh Krishna Computer Vision and Image Understanding https://www.sciencedirect.com/science/article/pii/S1077314223001753 |
| project page / code | https://github.com/tanmaysingha/SDBNetV2 |
| used Cityscapes data | fine annotations |
| used external data | |
| runtime | n/a |
| subsampling | no |
| submission date | July, 2022 |
| previous submissions |
Average results
| Metric | Value |
|---|---|
| IoU Classes | 72.362 |
| iIoU Classes | 43.8481 |
| IoU Categories | 87.8821 |
| iIoU Categories | 70.5684 |
Class results
| Class | IoU | iIoU |
|---|---|---|
| road | 98.1291 | - |
| sidewalk | 83.0438 | - |
| building | 91.5674 | - |
| wall | 53.337 | - |
| fence | 50.6664 | - |
| pole | 56.9552 | - |
| traffic light | 66.5632 | - |
| traffic sign | 71.369 | - |
| vegetation | 92.1769 | - |
| terrain | 68.5931 | - |
| sky | 94.7746 | - |
| person | 80.4948 | 56.0767 |
| rider | 61.3692 | 35.3794 |
| car | 93.8391 | 86.084 |
| truck | 58.2483 | 25.9516 |
| bus | 69.0523 | 36.2383 |
| train | 61.8443 | 31.8373 |
| motorcycle | 55.9914 | 29.4694 |
| bicycle | 66.8619 | 49.7484 |
Category results
| Category | IoU | iIoU |
|---|---|---|
| flat | 98.3339 | - |
| nature | 91.925 | - |
| object | 64.5255 | - |
| sky | 94.7746 | - |
| construction | 91.6965 | - |
| human | 80.7747 | 57.2908 |
| vehicle | 93.1445 | 83.8461 |