Valence-Arousal Estimation on AffectNet
0.61CCC (Valence)MT-VGG
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MT-VGGStrategy=Multi-Task, Backbone=VGG, Training Data=Aff-Wild22019.09 | 0.61 | 0.46 | |
| AlexNetBackbone=AlexNet2019.09 | 0.6 | 0.34 | |
| MT-Inc. ResNetStrategy=Multi-Task, Backbone=Inception ResNet, Training Data=Aff-Wild22019.09 | 0.52 | 0.45 | |
| ST-VGGStrategy=Single-Task, Backbone=VGG, Training Data=Aff-Wild22019.09 | 0.51 | 0.42 | |
| MT-SphereFaceStrategy=Multi-Task, Backbone=SphereFace, Training Data=Aff-Wild22019.09 | 0.5 | 0.43 | |
| ST-SphereFaceStrategy=Single-Task, Backbone=SphereFace, Training Data=Aff-Wild22019.09 | 0.5 | 0.4 | |
| ST-Inc. ResNetStrategy=Single-Task, Backbone=Inception ResNet, Training Data=Aff-Wild22019.09 | 0.5 | 0.42 |