Method Details
Details for method 'Qualcomm AI Research'
Method overview
| name | Qualcomm AI Research |
| challenge | pixel-level semantic labeling |
| details | Qualcomm AI Perception |
| publication | InverseForm: A Loss Function for Structured Boundary-Aware Segmentation https://arxiv.org/abs/2104.02745 |
| project page / code | |
| used Cityscapes data | fine annotations, coarse annotations |
| used external data | ImageNet |
| runtime | n/a |
| subsampling | no |
| submission date | May, 2021 |
| previous submissions | 1 |
Average results
| Metric | Value |
|---|---|
| IoU Classes | 85.6119 |
| iIoU Classes | 71.3567 |
| IoU Categories | 92.9927 |
| iIoU Categories | 85.6621 |
Class results
| Class | IoU | iIoU |
|---|---|---|
| road | 98.6017 | - |
| sidewalk | 89.4492 | - |
| building | 94.6114 | - |
| wall | 71.808 | - |
| fence | 69.2391 | - |
| pole | 75.838 | - |
| traffic light | 82.3792 | - |
| traffic sign | 85.532 | - |
| vegetation | 94.3274 | - |
| terrain | 74.9968 | - |
| sky | 95.8301 | - |
| person | 90.1758 | 78.8729 |
| rider | 79.7795 | 64.1955 |
| car | 96.9236 | 92.7701 |
| truck | 84.452 | 61.0671 |
| bus | 95.5841 | 72.1742 |
| train | 88.6489 | 67.1938 |
| motorcycle | 76.6814 | 62.4092 |
| bicycle | 81.7675 | 72.1707 |
Category results
| Category | IoU | iIoU |
|---|---|---|
| flat | 98.5743 | - |
| nature | 94.1197 | - |
| object | 80.9401 | - |
| sky | 95.8301 | - |
| construction | 94.6821 | - |
| human | 90.2216 | 79.733 |
| vehicle | 96.5814 | 91.5911 |