Hybrid Modeling of PhotoPlethySnyscography for noninvasive monitoring of cardiovascular parameters
Continuous heart monitoring can play an important role in intuitive health. However, some basic cardiac biomarkers of interest, including stroke volume and cardiac output, require invasive measurements, e.g. As a non-invasive alternative, PhotoPlethySyShography (PPG) measurements are routinely collected in hospital settings. Unfortunately, the prediction of key biomarkers from PPG instead of apw remains an open challenge, also complicated by the lack of descriptive PPG measurements. As a solution, we propose a hybrid ap- proach that uses hemodynamic masking and non-clinical data to measure cardiovascular biomarkers directly from PPG signals. Our Hybrid model combines a conditional autoencoder trained on patic-apw data with a stain density estimator for cardiac biomarkers trained on the generated apw components. As an important result, our experiments show that the proposed method can detect the variability of cardiac output and stroke volume and acpecform is a supervised basis for observing temporal changes in these biomalsers.
- † Eth Zurich, Switzerland
- ** Work Done while at Apple
