Technical Name Novel Pulse Wave Velocity (Vascular Age) Estimation System for Smart Watch
Project Operator National Taiwan University
Project Host 王宗道
Summary
Our team develops an accurate PWV estimation algorithm that uses wrist PPGECG signals from wearable devices. A missing-feature imputationambiguous-feature resolution technique is developedthe availability of wrist PPG morphological features is raised from 60 to 99.1. A weighted pulse decomposition approach is adopted5 component waves can be acquired to examine more detailed properties. The PWV is then estimated by XGBoost algorithm with the hierarchical regression model. The root mean square error of the estimation is reduced to 150 cm/s. The applicability of wearable device, especially smart watch, for long-termpervasive monitoring of personal health could be achieved by this technology.
Scientific Breakthrough
The morphology of wrist PPG is not as distinct as that of finger PPG signals. The proposed missing-feature imputationambiguous-feature resolution technique improves the availability of wrist PPG features from 60 to 99.1. With the weighted pulse decomposition approach, 5 component waves can be decomposed to acquire the latent forwardreflected waves so that the intrinsic hemodynamic properties can be realized. Based on these PPG features, the pulse wave velocity can be estimated by hierarchical regression model constructed by the XGBoost algorithm. The estimation root mean square error is 155 cm/s, showing significant improvement compared to the other published approachesfulfilling the standards set by the ARTERY society.
Industrial Applicability
The value of this technique can be analyzed in three aspects, the ICT industry, the medical application,the healthcare industry. With the estimation of pulse wave velocity (PWV), an integrative indicator of vascular health highly recommended by the medical society, smart watchessmart phones can upgrade its functionality, which is the trend for the development of personal long-term digital healthcare systems. In the medical field, it can enhance the diagnostic valuecan assist doctors for adjusting treatment strategies. Through the feedback of estimation for vascular age based on PWV, users can realize the effects of health management to improve his/her motivation.
Keyword Photoplethysmography (PPG) electrocardiography (ECG) wearable device smart watch pulse wave velocity feature extraction extreme gradient boost (XGBoost) weighted pulse decomposition vascular age regression
Notes
  • Contact
  • Claire Su
other people also saw