WESAD
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Activity Recognition | WESAD | Accuracy77 | 22 | |
| Affect recognition | WESAD (test) | Accuracy82.96 | 17 | |
| Emotion Distribution Learning | WESAD (subject-independent) | Chebyshev Distance0.0314 | 11 | |
| Emotion Distribution Learning | WESAD subject-independent (test) | Chebyshev Distance0.0314 | 11 | |
| Wearable stress and affect detection | WESAD | AUROC74.81 | 11 | |
| Classification | WESAD 10-shot | Accuracy81.8 | 10 | |
| Classification | WESAD 5-shot | Accuracy78.4 | 10 | |
| Classification | WESAD 3-shot | Accuracy77.9 | 10 | |
| Binary Classification (Low/High Arousal) | WESAD (Leave-one-participant-out (LOPO)) | Balanced Acc66 | 7 | |
| Binary Classification (Low/High Valence) | WESAD Time-Aware (TA) | Balanced Accuracy64 | 7 | |
| Cross-modal retrieval | WESAD | mAP30.04 | 6 | |
| Classification | WESAD | Accuracy76.18 | 6 | |
| Cross-modal knowledge transfer | WESAD ECG (Old) → PPG (New) (test) | BAcc49.57 | 6 | |
| Unsupervised cross-modal knowledge transfer | WESAD ECG -> PPG (test) | Balanced Accuracy49.57 | 6 | |
| FDI Detection | WESAD | Sensitivity (Sns)68.2 | 5 | |
| Human Sensing | WESAD 10-shot | Training Time (mins)4.33 | 5 | |
| Human Sensing | WESAD 5-shot | Training Time (mins)2.48 | 5 | |
| Human Sensing | WESAD 3-shot | Training Time (mins)1.49 | 5 | |
| Human Sensing | WESAD 10-shot | GPU Utilization77.52 | 5 | |
| Human Sensing | WESAD 5-shot | GPU Utilization70.27 | 5 | |
| Human Sensing | WESAD 3-shot | GPU Utilization52.58 | 5 | |
| Cross-modal knowledge transfer | WESAD PPG (Old) → ECG (New) (test) | BAcc62.96 | 5 | |
| Stress Detection | WESAD (5-fold stratified cross-val) | AUC99.46 | 4 | |
| Human Sensing | WESAD | Watch Latency (ms)460.5 | 4 | |
| Time-Series Segmentation | WESAD (test) | F-score64.1 | 4 |
Showing 25 of 29 rows