Pain Recognition on BioVid 77 source subjects 10 target subjects
97.11Subject 1 AccuracyFine-Tuning
Evaluation Results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Fine-TuningSetting=Oracle, Backbone=ResNet-182025.08 | 97.11 | 91.43 | 96.76 | 97.89 | 96.11 | 92.3 | 90.01 | 95.09 | 98.22 | 97 | 95.19 | |
| SFDA-PFTSetting=SFDA (data-based), Backbone=ResNet-182025.08 | 84.93 | 75.56 | 95.05 | 85.86 | 97.59 | 73.78 | 88.73 | 83.49 | 78.48 | 61.16 | 82.46 | |
| SFITSetting=SFDA (data-based), Backbone=ResNet-182025.08 | 80.92 | 68.79 | 82.87 | 84.7 | 91.62 | 57.32 | 60.59 | 74.56 | 60.71 | 79.94 | 74.2 | |
| DSFDASetting=SFDA (model-based), Backbone=ResNet-182025.08 | 77 | 75.11 | 90.89 | 85.67 | 88.11 | 67.48 | 89.22 | 87.56 | 69.22 | 72.11 | 80.24 | |
| SFDA-ITSetting=SFDA (data-based), Backbone=ResNet-182025.08 | 75.33 | 67.27 | 89 | 84.55 | 90.8 | 62.31 | 81.77 | 75.89 | 57.88 | 70.56 | 75.54 | |
| Source model (no adaptation)Setting=Source-only, Backbone=ResNet-182025.08 | 66.11 | 55.55 | 86.36 | 85.11 | 87.11 | 59 | 75.66 | 70.44 | 52.66 | 48.22 | 68.62 | |
| TPDSSetting=SFDA (model-based), Backbone=ResNet-182025.08 | 65.56 | 55.98 | 79.22 | 68.22 | 91.67 | 59.11 | 61.28 | 69.33 | 57.11 | 48.22 | 65.57 | |
| NRCSetting=SFDA (model-based), Backbone=ResNet-182025.08 | 59.33 | 39.38 | 84 | 72.89 | 79.67 | 42.67 | 63.92 | 53.95 | 54.89 | 42.47 | 60.31 | |
| SHOTSetting=SFDA (model-based), Backbone=ResNet-182025.08 | 55.78 | 47.76 | 41.05 | 52.44 | 54.67 | 48.89 | 54.46 | 49.22 | 54.86 | 44.44 | 50.35 |