Expression Recognition on Aff-Wild2 (val)
55.55F1 Score (P_EXPR)MAE+Transformer feature fusion
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MAE+Transformer feature fusionModality=Audio/video, Is ensemble?=Yes2026.03 | 55.55 | — | |
| EmotiEffNet + wav2vec 2.0, GLA, filtering + smoothingModality=Audio/video, Is ensemble?=No2026.03 | 47.4 | 57.98 | |
| CLIP+TCNModality=Video, Is ensemble?=Yes2026.03 | 46.51 | — | |
| EmotiEffNet + wav2vec 2.0, GLA, smoothingModality=Audio/video, Is ensemble?=No2026.03 | 46.04 | 57.5 | |
| EmotiEffNet, GLA, filtering + smoothingModality=Faces, Is ensemble?=No2026.03 | 45.79 | 55.69 | |
| EmotiEffNet, GLA, smoothingModality=Faces, Is ensemble?=No2026.03 | 44.85 | 55.41 | |
| Best EfficientNet + wav2vecModality=Audio/video, Is ensemble?=Yes2026.03 | 44.59 | 55.32 | |
| SSL + Temporal+ Post-processModality=Audio/video, Is ensemble?=Yes2026.03 | 44.43 | — | |
| wav2vec 2.0+EmotiEffNet, GLAModality=Audio/video, Is ensemble?=Yes2026.03 | 43.11 | 53.4 | |
| EmotiEffNet, GLA + filteringModality=Faces, Is ensemble?=No2026.03 | 42.5 | 52.22 | |
| V-NAWModality=Video, Is ensemble?=No2026.03 | 41.81 | — | |
| EmotiEffNet, GLAModality=Faces, Is ensemble?=No2026.03 | 41.4 | 51.62 | |
| EfficientNet-B0Modality=Faces, Is ensemble?=No2026.03 | 40.2 | — | |
| EmotiEffNet, focal loss, GLAModality=Faces, Is ensemble?=No2026.03 | 40.18 | 51.9 | |
| EmotiEffNet, focal lossModality=Faces, Is ensemble?=No2026.03 | 38.99 | 50.22 | |
| wav2vec 2.0, focal loss, GLA, smoothingModality=Audio, Is ensemble?=No2026.03 | 38.75 | 51.74 | |
| EmotiEffNetModality=Faces, Is ensemble?=No2026.03 | 38.68 | 47.59 | |
| CLIPModality=Faces, Is ensemble?=No2026.03 | 36 | — | |
| CLIP+LSTMModality=Video, Is ensemble?=Yes2026.03 | 34.1 | — | |
| wav2vec 2.0, focal loss, GLAModality=Audio, Is ensemble?=No2026.03 | 32.33 | 42.15 | |
| wav2vec 2.0, focal lossModality=Audio, Is ensemble?=No2026.03 | 30.5 | 40.08 | |
| Baseline VGGFACE (MixAugment)Modality=Faces, Is ensemble?=No2026.03 | 25 | — |