Video Instance Segmentation On Youtube Vis 3
評価指標
mAP_L
評価結果
このベンチマークにおける各モデルのパフォーマンス結果
モデル名 | mAP_L | Paper Title | Repository |
|---|---|---|---|
| CTVIS (ResNet-50) | 39.4 | CTVIS: Consistent Training for Online Video Instance Segmentation | |
| DVIS++(VIT-L) | 50.9 | DVIS++: Improved Decoupled Framework for Universal Video Segmentation | |
| DVIS(Swin-L) | 45.9 | DVIS: Decoupled Video Instance Segmentation Framework | |
| InstanceFormer (Swin) | 26.3 | InstanceFormer: An Online Video Instance Segmentation Framework | |
| CTVIS (Swin-L) | 46.4 | CTVIS: Consistent Training for Online Video Instance Segmentation | |
| InstanceFormer (Resnet-50) | 24.8 | InstanceFormer: An Online Video Instance Segmentation Framework | |
| CAVIS (VIT-L) | 48.6 | Context-Aware Video Instance Segmentation |
0 of 7 row(s) selected.