- Submitted by
- Oliver Zendel
- Authors
- Porzi, Lorenzo and Rota Bulò, Samuel and Colovic, Aleksander and Kontschieder, Peter
- Venue
- The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2019
- Abstract
- Seamless Scene Segmentation Resnet101, pretrained on Imagenet; supplied with altered MVD to include WildDash2 classes; does not contain other RVC label policies (i.e. no ADE20K/COCO-specific classes -> rvcsubset and not a proper submission); Trained and Submitted by Oliver Zendel for RVC
- Runtime
- 500 ms
- External data used
- model pre-trained on ImageNet; regular training on altered MVD to contain WildDash2 labels (van, pickup) (20k frames) no other dataset (i.e. no VIPER frames)
- Links
-
- Number of submissions
- 1
- Detailed results
-
| Class |
PQ (mean) |
SQ (mean) |
RQ (mean) |
| Sky | 66.3 | 86.3 | 75.5 |
| Road | 54.7 | 81.5 | 66.9 |
| Sidewalk | 37.3 | 71.6 | 51.7 |
| Terrain | 14.9 | 66.5 | 21.9 |
| Tree | 0.0 | 0.0 | 0.0 |
| Vegetation | 2.2 | 63.1 | 3.6 |
| Building | 34.4 | 72.7 | 47.2 |
| Infrastructure | 0.0 | 0.0 | 0.0 |
| Fence | 2.3 | 65.5 | 3.5 |
| Billboard | 16.1 | 68.3 | 23.6 |
| Trafficlight | 5.2 | 65.9 | 7.9 |
| Trafficsign | 34.9 | 79.2 | 44.0 |
| Mobilebarrier | 0.0 | 0.0 | 0.0 |
| Firehydrant | 17.3 | 78.8 | 21.9 |
| Chair | 0.0 | 0.0 | 0.0 |
| Trash | 0.0 | 0.0 | 0.0 |
| Trashcan | 9.8 | 77.1 | 12.7 |
| Person | 13.6 | 72.2 | 18.7 |
| Motorcycle | 18.8 | 66.1 | 28.5 |
| Car | 28.5 | 78.1 | 36.4 |
| Van | 11.0 | 65.4 | 13.6 |
| Bus | 17.9 | 82.0 | 21.9 |
| Truck | 26.8 | 81.3 | 32.9 |
| Stuff | 20.2 | 50.4 | 26.0 |
| Things | 14.9 | 66.7 | 19.4 |
| Mean | 17.9 | 57.5 | 23.1 |