- Submitted by
- Xingyi Zhou
- Authors
- Xingyi Zhou
- Venue
- UniDet_RVC
- Abstract
- Learning a unified label space from data
- Runtime
- 300 ms
- External data used
- COCO, Cityscapes, OpenImages, Mapillary, ScanNet
- Links
-
- There is no link to the paper.
- There is no link to the code.
- There is no link to the project website.
- Number of submissions
- 2
- Detailed results
-
| Class |
AP (day) |
AP (sunset) |
AP (rain) |
AP (snow) |
AP (night) |
AP (mean) |
| Trafficlight | 40.5 | 35.5 | 35.1 | 47.4 | 3.1 | 32.3 |
| Firehydrant | 28.8 | 25.3 | 39.1 | 32.7 | 7.5 | 26.7 |
| Chair | 2.9 | 0.8 | 1.4 | 2.6 | 0.1 | 1.6 |
| Trashcan | 21.9 | 13.2 | 16.4 | 16.5 | 1.5 | 13.9 |
| Person | 19.0 | 11.9 | 7.6 | 18.4 | 0.5 | 11.5 |
| Motorcycle | 20.8 | 15.2 | 14.0 | 12.6 | 5.1 | 13.5 |
| Car | 35.2 | 34.6 | 39.5 | 40.5 | 6.6 | 31.3 |
| Van | 27.3 | 3.1 | 20.8 | 7.4 | 18.6 | 15.4 |
| Bus | 23.7 | 21.4 | 26.1 | 15.9 | 0.5 | 17.5 |
| Truck | 30.9 | 21.2 | 29.6 | 28.2 | 16.7 | 25.3 |
| Mean | 25.1 | 18.2 | 23.0 | 22.2 | 6.0 | 18.9 |
| Class |
AP50 (day) |
AP50 (sunset) |
AP50 (rain) |
AP50 (snow) |
AP50 (night) |
AP50 (mean) |
| Trafficlight | 69.1 | 64.9 | 63.4 | 73.6 | 5.9 | 55.4 |
| Firehydrant | 59.6 | 50.2 | 78.6 | 57.4 | 16.3 | 52.4 |
| Chair | 9.2 | 2.4 | 6.2 | 7.2 | 0.3 | 5.0 |
| Trashcan | 40.5 | 24.8 | 33.6 | 29.9 | 2.9 | 26.3 |
| Person | 47.2 | 33.5 | 21.5 | 48.8 | 1.7 | 30.5 |
| Motorcycle | 52.2 | 56.9 | 35.0 | 46.0 | 14.0 | 40.8 |
| Car | 63.1 | 62.5 | 65.9 | 64.9 | 11.3 | 53.5 |
| Van | 38.5 | 4.8 | 31.9 | 10.3 | 24.4 | 22.0 |
| Bus | 41.6 | 37.8 | 34.5 | 24.9 | 0.9 | 27.9 |
| Truck | 52.0 | 40.0 | 48.0 | 42.9 | 25.6 | 41.7 |
| Mean | 47.3 | 37.8 | 41.9 | 40.6 | 10.3 | 35.6 |