Details for method 'GoogLeNetV1_ROB'
Method overview
| name |
GoogLeNetV1_ROB |
| challenge |
pixel-level semantic labeling |
| details |
GoogLeNet-v1 FCN trained on Cityscapes, KITTI, and ScanNet, as required by the Robust Vision Challenge at CVPR'18 (http://robustvision.net/) |
| publication |
Anonymous
|
| project page / code |
|
| used Cityscapes data |
fine annotations |
| used external data |
ImageNet, KITTI, ScanNet |
| runtime |
n/a |
| subsampling |
no |
| submission date |
April, 2018 |
| previous submissions |
|
Average results
| Metric |
Value |
| IoU Classes | 59.6107 |
| iIoU Classes | 35.1399 |
| IoU Categories | 83.0324 |
| iIoU Categories | 64.4251 |
Class results
| Class |
IoU |
iIoU |
| road | 96.6092 | - |
| sidewalk | 73.7704 | - |
| building | 87.1435 | - |
| wall | 27.116 | - |
| fence | 31.5918 | - |
| pole | 47.2207 | - |
| traffic light | 53.2359 | - |
| traffic sign | 59.0071 | - |
| vegetation | 89.6268 | - |
| terrain | 55.0949 | - |
| sky | 92.1503 | - |
| person | 72.3161 | 46.8655 |
| rider | 48.2823 | 27.2006 |
| car | 90.8773 | 83.4367 |
| truck | 29.773 | 16.1183 |
| bus | 40.0204 | 24.4205 |
| train | 33.8484 | 18.2762 |
| motorcycle | 42.9019 | 22.6491 |
| bicycle | 62.0171 | 42.1522 |
Category results
| Category |
IoU |
iIoU |
| flat | 97.3446 | - |
| nature | 88.4961 | - |
| object | 53.8987 | - |
| sky | 92.1503 | - |
| construction | 87.3971 | - |
| human | 73.5065 | 48.5833 |
| vehicle | 88.4336 | 80.2669 |
Links
Download results as .csv file
Benchmark page
