Audio-Visual Segmentation on AVSBench-object S4
81.2mIoUTB-AVA
Evaluation results
| Method | Links | |
|---|---|---|
| TB-AVAVisual backbone=SigLIP2-L/16, Audio backbone=BEATs2026.05 | 81.2 | |
| AVMoEVisual backbone=Swin-V2-L, Audio backbone=HTS-AT2026.05 | 81.1 | |
| DG-SCTVisual backbone=Swin-V2-L, Audio backbone=HTS-AT2026.05 | 80.9 | |
| MoLTVisual backbone=Swin-V2-L, Audio backbone=HTS-AT2026.05 | 80.8 | |
| MettleVisual backbone=Swin-V2-L, Audio backbone=HTS-AT2026.05 | 80.7 | |
| LAVisHVisual backbone=Swin-V2-L (shared), Audio backbone=Swin-V2-L (shared)2026.05 | 80.1 | |
| AVSVisual backbone=PVT-v2, Audio backbone=VGGish2026.05 | 78.7 |