Instance Segmentation on KITTI (val)
16.9mAP (Mask)UniDet
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| UniDetTraining Dataset=Unified (COCO, CityScapes, Mapillary, VIPER, ScanNet, OpenImages)2021.02 | 16.9 | — | — | — | |
| COCO-trained modelTraining Dataset=COCO2021.02 | 15.7 | — | — | — | |
| Mapillary-trained modelTraining Dataset=Mapillary2021.02 | 13.4 | — | — | — | |
| CityScapes-trained modelTraining Dataset=CityScapes2021.02 | 13 | — | — | — | |
| OpenImages-trained modelTraining Dataset=OpenImages2021.02 | 7.2 | — | — | — | |
| VIPER-trained modelTraining Dataset=VIPER2021.02 | 6.5 | — | — | — | |
| ScanNet-trained modelTraining Dataset=ScanNet2021.02 | 0 | — | — | — | |
| SqueezeSegCategory=car, CRF=included2017.10 | — | 63.4 | 90.7 | 59.5 | |
| SqueezeSegCategory=car, CRF=excluded2017.10 | — | 60 | 91.3 | 56.7 | |
| SqueezeSegCategory=pedestrian, CRF=included2017.10 | — | 43.5 | 28.6 | 20.8 | |
| SqueezeSegCategory=pedestrian, CRF=excluded2017.10 | — | 50.8 | 27.5 | 21.7 | |
| SqueezeSegCategory=cyclist, CRF=included2017.10 | — | 30.4 | 39 | 20.6 | |
| SqueezeSegCategory=cyclist, CRF=excluded2017.10 | — | 30.1 | 43.7 | 21.7 |