Unsupervised Multi-object Segmentation on KITTI
58.3FG-ARIPPMP (Swin + WL)
Evaluation results
| Method | Links | |
|---|---|---|
| PPMP (Swin + WL)Backbone=Swin, Optical flow estimation model=RAFT, Evaluation resolution=96 x 320, Warp loss (WL)=true2022.10 | 58.3 | |
| PPMP (WL)Backbone=ResNet-18, Optical flow estimation model=RAFT, Evaluation resolution=96 x 320, Warp loss (WL)=true2022.10 | 51.9 | |
| PPMPBackbone=ResNet-18, Optical flow estimation model=RAFT, Evaluation resolution=96 x 3202022.10 | 50.8 | |
| Bao et al.Backbone=ResNet-18, Optical flow estimation model=RAFT2022.10 | 47.1 | |
| MCGBackbone=ResNet-182022.10 | 40.9 | |
| SCALORBackbone=ResNet-182022.10 | 21.1 | |
| MONetBackbone=ResNet-182022.10 | 14.9 | |
| S-IODINEBackbone=ResNet-182022.10 | 14.4 | |
| SABackbone=ResNet-182022.10 | 13.8 |