Audio-Guided Video Object Segmentation on MeViS v2
42.3J&F ScoreLMPM++
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| LMPM++2025.12 | 42.3 | 39.1 | 45.5 | 43.2 | 85.4 | |
| LMPM_GT TextInput modality=Ground Truth Text2025.12 | 38.3 | 35.6 | 40.9 | 20.1 | 72.9 | |
| LMPM_AudioInput modality=Audio features via Whisper2025.12 | 37.5 | 34.5 | 40.4 | 11 | 78.8 | |
| LMPM_Audio→TextInput modality=Audio converted to text via Whisper-Base2025.12 | 36.8 | 33.2 | 40.5 | 22.1 | 64.7 | |
| MUTR2025.12 | 33.6 | 30.9 | 36.3 | 0 | 100 | |
| WNet2025.12 | 16.5 | 16.6 | 16.3 | 0.5 | 99.3 |