Audio-Visual Segmentation on AVS (test)
81.8ScoreSTG-CMA
Evaluation Results
| Method | Links | |
|---|---|---|
| STG-CMABackbone (Vision/Audio)=Swin-L (Shared), Param Update=38.6, Total Parameters=233.6, Update Ratio=16.5%2026.04 | 81.8 | |
| DG-SCTBackbone (Vision/Audio)=Swin-L/HTS-AT, Param Update=61.5, Total Parameters=594.8, Update Ratio=10.3%2026.04 | 80.9 | |
| CoLABackbone (Vision/Audio)=Swin-L/SSLAM, Param Update=44.8, Total Parameters=364.0, Update Ratio=12.3%2026.04 | 80.9 | |
| LAVisHBackbone (Vision/Audio)=Swin-L (Shared), Param Update=37.2, Total Parameters=266.4, Update Ratio=14.0%2026.04 | 80.1 |