Gesture-to-Speech Synthesis on PATS Same Text (test)
0.9471Gesture OffsetGesture2Speech
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Gesture2Speechvariant=multimodal MoE2026.03 | 0.9471 | 0.0559 | 17.55 | 12.14 | 3.7 | 3.65 | 3.19 | |
| Gesture2Speechvariant=unimodal MoE2026.03 | 0.9794 | 0.0404 | 22.42 | 15.2 | 3.44 | 3.45 | 3.12 | |
| Ground Truth2026.03 | 1.0198 | 0.0362 | 35.61 | 25.2 | 3.34 | 3.32 | — | |
| Gesture2Speechvariant=XTTS V22026.03 | 1.0386 | 0.0382 | 20.27 | 14.85 | 3.34 | 3.34 | 3.08 | |
| Gesture2Speechvariant=H-MoE2026.03 | 1.2008 | 0.0357 | 16.93 | 11.74 | 3.46 | 3.36 | 3.12 | |
| Gesture2Speechvariant=GPT-SoVITS2026.03 | 1.8656 | 0.007 | 34.04 | 26 | 3.17 | 3.51 | 3.14 |