Vocal Tract Segmentation on USC-TIMIT (US-UT)
83.78DSC (%)Ours
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| OursMethod Type=Multimodal2026.05 | 83.78 | 3.43 | |
| AV-SAMMethod Type=Multimodal, Adapted=true2026.05 | 80.62 | 4.26 | |
| Jain et al.Method Type=Multimodal2026.05 | 80.14 | 3.5 | |
| U-MambaMethod Type=Unimodal2026.05 | 78.06 | 4.11 | |
| AVSegFormerMethod Type=Multimodal, Adapted=true2026.05 | 76.59 | 3.02 | |
| DINOv2Method Type=Unimodal2026.05 | 71.38 | 4.42 | |
| Swin-UNETRMethod Type=Unimodal2026.05 | 69.4 | 4.86 | |
| MedSAM-2Method Type=Unimodal2026.05 | 69.22 | 4.48 | |
| nnU-NetMethod Type=Unimodal2026.05 | 66.55 | 7.81 | |
| ResUNet-aMethod Type=Unimodal2026.05 | 65.67 | 7.66 | |
| MedSegDiff-V2Method Type=Unimodal2026.05 | 63.64 | 5.59 |