Video Panoptic Segmentation On Cityscapes Vps
評価指標
VPQ
VPQ (stuff)
VPQ (thing)
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 | VPQ | VPQ (stuff) | VPQ (thing) | Paper Title | Repository |
|---|---|---|---|---|---|
| TarViS (Swin-T) | 58.0 | 69.0 | 42.9 | TarViS: A Unified Approach for Target-based Video Segmentation | |
| TarViS (ResNet-50) | 53.3 | 66.0 | 35.9 | TarViS: A Unified Approach for Target-based Video Segmentation | |
| PolyphonicFormer | 62.3 | - | - | PolyphonicFormer: Unified Query Learning for Depth-aware Video Panoptic Segmentation | |
| Video K-Net (Swin-B) | 62.2 | 71.8 | 49.8 | Video K-Net: A Simple, Strong, and Unified Baseline for Video Segmentation | |
| VIP-Deeplab | 63.1 | 73.0 | 49.5 | ViP-DeepLab: Learning Visual Perception with Depth-aware Video Panoptic Segmentation | |
| TarViS (Swin-L) | 58.9 | 69.9 | 43.7 | TarViS: A Unified Approach for Target-based Video Segmentation | |
| VPSNet-SiamTrack | 57.3 | 66.4 | 44.7 | Learning to Associate Every Segment for Video Panoptic Segmentation | - |
| VPSNet | 57.0 | 66.0 | 44.7 | Video Panoptic Segmentation |
0 of 8 row(s) selected.