Categorical Emotion Recognition on MAFW 11-class
34.24UAREmoCLIP (Frozen backbone)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| EmoCLIP (Frozen backbone)Mode=Supervised, Architecture=EmoCLIP (Frozen backbone), Contrastive Pre-training=MAFW [class descriptions]2023.10 | 34.24 | 41.46 | |
| T-ESFLMode=Supervised, Architecture=T-ESFL2023.10 | 33.28 | 48.18 | |
| VIT_LSTMMode=Supervised, Architecture=VIT_LSTM2023.10 | 32.67 | 45.56 | |
| C3DMode=Supervised, Architecture=C3D2023.10 | 31.17 | 42.25 | |
| EmoCLIP (LP)Mode=Supervised, Architecture=EmoCLIP (LP)2023.10 | 30.26 | 44.23 | |
| C3D_LSTMMode=Supervised, Architecture=C3D_LSTM2023.10 | 29.75 | 43.76 | |
| Resnet18_LSTMMode=Supervised, Architecture=Resnet18_LSTM2023.10 | 28.08 | 39.38 | |
| EmoCLIPMode=Zero-shot, Architecture=EmoCLIP, Contrastive Pre-training=MAFW [sample descriptions]2023.10 | 25.86 | 33.49 | |
| CLIPMode=Zero-shot, Architecture=CLIP, Contrastive Pre-training=Laion-400m2023.10 | 20.4 | 21.16 | |
| FaRL - ViT-B/16Mode=Zero-shot, Architecture=FaRL - ViT-B/16, Contrastive Pre-training=Laion Face-20M2023.10 | 14.07 | 7.7 | |
| VAEmoSSL=true, Modality=A+V, #Params(M)=392025.05 | 0.4567 | 0.5891 | |
| FineCLIPERSSL=true, Modality=T+V, #Params(M)=N/A2025.05 | 0.4501 | 0.5691 | |
| MMA-DFERSSL=true, Modality=A+V2025.05 | 0.4425 | 0.5845 | |
| S4DSSL=true, Modality=V, #Params(M)=1012025.05 | 0.4372 | 0.5844 | |
| HICMAE-BSSL=true, Modality=A+V, #Params(M)=812025.05 | 0.4265 | 0.5617 | |
| HICMAE-SSSL=true, Modality=A+V, #Params(M)=462025.05 | 0.4166 | 0.5445 | |
| MAE-DFERSSL=true, Modality=V, #Params(M)=852025.05 | 0.4162 | 0.5431 | |
| SVFAPSSL=true, Modality=V, #Params(M)=782025.05 | 0.4119 | 0.5428 | |
| DFER-CLIPSSL=false, Modality=V, #Params(M)=1532025.05 | 0.3989 | 0.5255 | |
| T-MEPSSL=false, Modality=A+V, #Params(M)=612025.05 | 0.3717 | 0.5115 | |
| T-ESFLSSL=false, Modality=A+V2025.05 | 0.3335 | 0.487 | |
| WavLM-PlusSSL=true, Modality=A, #Params(M)=952025.05 | 0.2633 | 0.3407 | |
| HUBERTSSL=true, Modality=A, #Params(M)=952025.05 | 0.25 | 0.326 |