Improved Biometric Stress Monitoring Using Heart Rate Variability and Capsule Networks
Author Information
Author(s): M. Khayyat, Munshi Raafat M., Alabduallah Bayan, Lamoudan Tarik, Ghith Ehab, Kim Tai-hoon, A. Abdelhamid Abdelaziz
Primary Institution: University of Jeddah, Saudi Arabia
Hypothesis
Can a Capsule Network model improve the accuracy of biometric stress monitoring using heart rate variability data?
Conclusion
The proposed Capsule Network model achieved high accuracy in monitoring stress levels, outperforming traditional models.
Supporting Evidence
- The Capsule Network model achieved an accuracy of 92.76% on the Swell dataset.
- On the WESAD dataset, the model reached an accuracy of 96.76%.
- The binary classification accuracy for stress detection was 98.52% on the Swell dataset and 99.82% on the WESAD dataset.
- 5-fold cross-validation confirmed the robustness of the model.
Takeaway
This study shows that using advanced technology can help us better understand and monitor stress in people, especially at work.
Methodology
The study utilized heart rate variability data and a Capsule Network model to monitor stress levels, validated through extensive experiments on two datasets.
Limitations
The study may be limited by the datasets used and the generalizability of the findings to broader populations.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website