Motion Segmentation on FBMS59 (val)
75.6J-measureCOSNet
Evaluation results
| Method | Links | |
|---|---|---|
| COSNetTraining=Supervised, Input=Image2022.01 | 75.6 | |
| DyStab - Stat&DynTraining=Supervised Features, Input=Image & Flow2022.01 | 74.2 | |
| CIS - With PostTraining=Fully Unsupervised, Input=Image & Flow2022.01 | 63.6 | |
| ARPTraining=Supervised Features, Input=Image & Flow2022.01 | 59.8 | |
| EM-driven networkTraining=Fully Unsupervised, Input=Flow2022.01 | 57.8 | |
| MoSegTraining=Fully Unsupervised, Input=Flow2022.01 | 53.1 | |
| DyStab - DynTraining=Supervised Features, Input=Flow2022.01 | 49.1 | |
| FTSTraining=Fully Unsupervised, Input=Flow2022.01 | 47.7 | |
| CIS - No PostTraining=Fully Unsupervised, Input=Image & Flow2022.01 | 36.8 |