Method Details
Details for method 'CGMANet_v1'
Method overview
| name | CGMANet_v1 |
| challenge | pixel-level semantic labeling |
| details | Context Guided Multi-scale Attention for Real-time Semantic Segmentation of Road-scene |
| publication | Context Guided Multi-scale Attention for Real-time Semantic Segmentation of Road-scene Saquib Mazhar |
| project page / code | |
| used Cityscapes data | fine annotations |
| used external data | |
| runtime | n/a |
| subsampling | no |
| submission date | June, 2024 |
| previous submissions |
Average results
| Metric | Value |
|---|---|
| IoU Classes | 73.3381 |
| iIoU Classes | 48.2996 |
| IoU Categories | 88.4921 |
| iIoU Categories | 74.854 |
Class results
| Class | IoU | iIoU |
|---|---|---|
| road | 97.7884 | - |
| sidewalk | 82.1103 | - |
| building | 91.64 | - |
| wall | 52.0655 | - |
| fence | 49.8619 | - |
| pole | 61.8218 | - |
| traffic light | 68.3498 | - |
| traffic sign | 72.5671 | - |
| vegetation | 91.8948 | - |
| terrain | 68.4075 | - |
| sky | 94.2718 | - |
| person | 82.7412 | 63.7015 |
| rider | 63.719 | 40.06 |
| car | 93.9462 | 86.8755 |
| truck | 56.3505 | 26.4804 |
| bus | 75.3023 | 41.7619 |
| train | 60.7609 | 34.6774 |
| motorcycle | 59.0205 | 36.8576 |
| bicycle | 70.8053 | 55.9823 |
Category results
| Category | IoU | iIoU |
|---|---|---|
| flat | 97.7465 | - |
| nature | 91.5373 | - |
| object | 68.1057 | - |
| sky | 94.2718 | - |
| construction | 91.6021 | - |
| human | 82.9205 | 64.925 |
| vehicle | 93.2606 | 84.7831 |