Valence-Arousal Estimation on AFEW-VA (test)
0.643CCC-VELIM
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| ELIMCase=Static, Backbone=Mlp-Mixer2022.09 | 0.643 | 0.598 | 0.167 | 0.164 | 0.747 | 0.877 | 0.704 | 0.707 | |
| ELIMCase=Static, Backbone=ResNet-182022.09 | 0.615 | 0.614 | 0.168 | 0.17 | 0.723 | 0.876 | 0.651 | 0.679 | |
| ELIMCase=Static, Tuning Strategy=AL-tuned2022.09 | 0.602 | 0.581 | 0.186 | 0.198 | 0.692 | 0.815 | 0.68 | 0.645 | |
| CAFCase=Static, Backbone=ResNet-182022.09 | 0.59 | 0.54 | 0.17 | 0.18 | 0.68 | 0.87 | 0.67 | 0.6 | |
| CAFCase=Static, Tuning Strategy=AL-tuned2022.09 | 0.58 | 0.57 | 0.2 | 0.2 | 0.66 | 0.83 | 0.67 | 0.63 | |
| FaceBehaviorNettraining=both coupling losses, post-processing=true2021.05 | 0.57 | 0.59 | — | — | — | — | — | — | |
| HO-ConvCase=Temporal, Configuration=trans2022.09 | 0.57 | 0.56 | 0.2 | 0.21 | 0.67 | 0.79 | 0.64 | 0.62 | |
| FaceBehaviorNettraining=both coupling losses, post-processing=false2021.05 | 0.55 | 0.57 | — | — | — | — | — | — | |
| Kossaifi et al.Case=Static2022.09 | 0.55 | 0.52 | 0.24 | 0.24 | 0.64 | 0.77 | 0.55 | 0.57 | |
| AffWildNetstatus=fine-tuned2021.05 | 0.52 | 0.56 | — | — | — | — | — | — | |
| Kollias et al.Case=Temporal2022.09 | 0.52 | 0.56 | — | — | — | — | 0.51 | 0.58 | |
| Mitenkova et al.Case=Static2022.09 | 0.33 | 0.4 | 0.4 | 0.41 | — | — | 0.33 | 0.42 | |
| HO-ConvCase=Temporal2022.09 | 0.11 | 0.15 | 0.28 | 0.19 | 0.53 | 0.75 | 0.12 | 0.23 |