Facial Expression Recognition on BioVid (target subjects (10))
92.22Accuracy (Sub-1)Supervised fine-tuning
Evaluation results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Supervised fine-tuningSetting=Oracle2025.08 | 92.22 | 86.83 | 91.89 | 92.96 | 91.27 | 87.65 | 85.48 | 90.3 | 93.28 | 92.12 | 90.4 | |
| SFDA-PFTSetting=SFDA (data-based), Backbone=ResNet-18, Optimizer=Adam, Learning rate=1 x 10^-3, Batch size=64, Learning rate scheduler=ReduceLROnPlateau, lambda_expr=1.0, lambda_style=0.32025.08 | 80.65 | 71.75 | 90.26 | 81.54 | 92.68 | 70.06 | 84.26 | 79.29 | 74.53 | 58.08 | 78.31 | |
| SFITSetting=SFDA (data-based)2025.08 | 76.85 | 65.33 | 78.7 | 80.44 | 87.01 | 54.44 | 57.54 | 70.81 | 57.66 | 75.92 | 70.47 | |
| SFDA-ITSetting=SFDA (data-based)2025.08 | 71.54 | 63.89 | 84.53 | 80.3 | 86.24 | 59.18 | 77.66 | 72.08 | 54.97 | 67.01 | 71.74 | |
| DSFDASetting=SFDA (model-based)2025.08 | 65.72 | 64.1 | 77.57 | 73.12 | 75.2 | 57.59 | 76.15 | 74.73 | 59.08 | 61.54 | 68.48 | |
| Source model (no adaptation)Setting=Source-only2025.08 | 62.78 | 52.76 | 82.02 | 80.83 | 82.73 | 56.03 | 71.85 | 66.9 | 50.01 | 45.79 | 65.17 | |
| TPDSSetting=SFDA (model-based)2025.08 | 62.26 | 53.16 | 75.23 | 64.79 | 87.06 | 56.14 | 58.2 | 65.84 | 54.24 | 45.79 | 62.27 | |
| SHOTSetting=SFDA (model-based)2025.08 | 52.97 | 45.35 | 38.98 | 49.8 | 51.92 | 46.43 | 51.72 | 46.74 | 52.1 | 42.2 | 47.82 | |
| NRCSetting=SFDA (model-based)2025.08 | 48.45 | 32.16 | 68.6 | 59.52 | 65.06 | 34.85 | 52.2 | 44.06 | 44.82 | 34.68 | 48.44 |