Audio-visual Speech Recognition on LRS3 Pixelated face
26WER (Babble, -10 dB)
CAV2vec
Updated 3d ago
Evaluation Results
| Method | Links | |||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2025.12 | 26 | 12 | 4.7 | 2.5 | 1.9 | 9.4 | 5.8 | 3.6 | 2.5 | 2.3 | 1.7 | 3.2 | 9.6 | 4.2 | 2.6 | 1.9 | 1.7 | 4 | 5.1 | 7.3 | 1.5 | |
2025.12 | 29.8 | 13.6 | 5 | 2.6 | 1.9 | 10.6 | 7.4 | 4.1 | 2.7 | 2.1 | 2 | 3.7 | 10.9 | 5.1 | 2.8 | 2.2 | 1.9 | 4.6 | 5.8 | 8.3 | 1.6 | |
2025.12 | 30.1 | 13.1 | 5.2 | 2.6 | 2 | 10.6 | 7.3 | 4.3 | 3 | 2.2 | 1.9 | 3.7 | 10.3 | 5 | 3.1 | 2.1 | 1.8 | 4.5 | 5.8 | 8.3 | 1.5 | |
2025.12 | 30.6 | 14.3 | 5.2 | 2.6 | 2 | 10.9 | 7.5 | 4.5 | 3 | 2.3 | 2 | 3.9 | 10.7 | 5.4 | 2.9 | 2.1 | 1.7 | 4.6 | 6 | 8.6 | 1.5 | |
2025.12 | 39.1 | 21.6 | 5.7 | 2.7 | 1.9 | 14.2 | 42.4 | 31 | 13.3 | 4.8 | 2.2 | 18.7 | 18.5 | 7.6 | 3.4 | 2.2 | 1.9 | 6.7 | 11.6 | 17.7 | 1.7 | |
2025.12 | 43.7 | 25.2 | 8.5 | 3.6 | 2.8 | 16.8 | 60 | 42.8 | 18.2 | 5 | 3.2 | 25.8 | 23.4 | 10.9 | 5.2 | 3.2 | 2.6 | 9.1 | 15.2 | 23.1 | 2.3 | |
2025.12 | 55.4 | 31.8 | 13.7 | 7.5 | 5.3 | 22.7 | 44.4 | 28.3 | 16.8 | 10.1 | 7 | 21.3 | 31.8 | 17.9 | 9.4 | 6.4 | 5.2 | 14.1 | 18.1 | 25.7 | 4.2 |