Schmidt 2018 — WESAD: Wearable Stress and Affect Detection
Citation: Schmidt, P., Reiss, A., Duerichen, R., et al. (2018). Proceedings of the 2018 ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp 2018).
DOI: 10.1145/3242969.3242985
Type: Dataset paper
Key Contribution
Introduced the most widely used public dataset for wearable stress detection. Dual-device (wrist + chest) protocol with validated stress induction.
Design
- 15 subjects, lab-based
- Trier Social Stress Test (public speaking + mental arithmetic)
- Amusement condition (funny video)
- Meditation (guided breathing)
- Two wearable devices simultaneously
Key Results
- Chest-worn: 3-class ~93%, 2-class ~93%
- Wrist-worn (E4): 3-class ~75%, 2-class ~87%
- Random Forest best overall
Impact
300+ citations. Standard benchmark for wearable affect recognition. Enabled the entire field of wrist-based stress detection research.
Linked To
- Dataset: wesad
- Dataset: feel-benchmark (WESAD included)