Method Details
Details for method 'DeepLab LargeFOV StrongWeak'
Method overview
| name | DeepLab LargeFOV StrongWeak |
| challenge | pixel-level semantic labeling |
| details | Trained on a pre-release version of the dataset |
| publication | Weakly- and Semi-Supervised Learning of a DCNN for Semantic Image Segmentation G. Papandreou, L.-C. Chen, K. Murphy, and A. L. Yuille ICCV 2015 http://www.cv-foundation.org/openaccess/content_iccv_2015/papers/Papandreou_Weakly-_and_Semi-Supervised_ICCV_2015_paper.pdf |
| project page / code | https://bitbucket.org/deeplab/deeplab-public/ |
| used Cityscapes data | fine annotations, coarse annotations |
| used external data | ImageNet |
| runtime | 4 s |
| subsampling | 2 |
| submission date | April, 2016 |
| previous submissions |
Average results
| Metric | Value |
|---|---|
| IoU Classes | 64.7945 |
| iIoU Classes | 34.9024 |
| IoU Categories | 81.2912 |
| iIoU Categories | 58.6522 |
Class results
| Class | IoU | iIoU |
|---|---|---|
| road | 97.3509 | - |
| sidewalk | 78.2755 | - |
| building | 88.0623 | - |
| wall | 47.4569 | - |
| fence | 44.1554 | - |
| pole | 29.4742 | - |
| traffic light | 44.3612 | - |
| traffic sign | 55.4453 | - |
| vegetation | 89.3685 | - |
| terrain | 67.3123 | - |
| sky | 92.8087 | - |
| person | 71.0044 | 40.6891 |
| rider | 49.2612 | 23.0572 |
| car | 91.4369 | 78.6038 |
| truck | 55.9035 | 21.397 |
| bus | 66.5594 | 32.3992 |
| train | 56.6897 | 27.5993 |
| motorcycle | 48.0682 | 20.8278 |
| bicycle | 58.1014 | 34.6458 |
Category results
| Category | IoU | iIoU |
|---|---|---|
| flat | 97.7611 | - |
| nature | 89.0106 | - |
| object | 40.448 | - |
| sky | 92.8087 | - |
| construction | 88.187 | - |
| human | 70.87 | 41.4042 |
| vehicle | 89.9533 | 75.9002 |