Video Object Segmentation on DAVIS (Gth Metric)
86.2GthXMem
Evaluation Results
| Method | Links | |
|---|---|---|
| XMemBackbone=R502024.02 | 86.2 | |
| DeAOTBackbone=Swin-B2024.02 | 86.2 | |
| TarVISBackbone=SwinL, Joint Training=false, Universal=true2024.02 | 85.2 | |
| DeAOTBackbone=R502024.02 | 85.2 | |
| TarVISBackbone=ResNet50, Joint Training=false, Universal=true2024.02 | 82.6 | |
| UNINEXTBackbone=ViT-H, Joint Training=false, Universal=false2024.02 | 81.8 | |
| UNINEXTBackbone=ConvNeXtL, Joint Training=false, Universal=false2024.02 | 77.2 | |
| UniVSBackbone=SwinL, Joint Training=true, Universal=true2024.02 | 76.2 | |
| UniVSBackbone=SwinB, Joint Training=true, Universal=true2024.02 | 75 | |
| UNINEXTBackbone=ResNet50, Joint Training=false, Universal=false2024.02 | 74.5 | |
| UniVSBackbone=SwinT, Joint Training=true, Universal=true2024.02 | 71.7 | |
| UniVSBackbone=ResNet50, Joint Training=true, Universal=true2024.02 | 70.5 |