Photoplethysmography Wave Decomposition by Machine Learning
*Yu-ching Hung (National Cheng Kung University (NCKU)) firstname.lastname@example.org
In this talk, I will present an optimization problem about the photoplethysmography (PPG) signals recorded at fingertip. We assume that the recorded wave form is composed by the
wave from heart and its reflection wave from the branch of main artery. Since we don't have the original wave from heart, we use some constructive waves such as Gaussian signal and
Poisson signal instead. The time delay and amplitude of the reflection wave are computed from minimizing the least square error of the constructive wave and the recorded wave.
The time delay and amplitude of the reflection wave are related to the stiffness index and reflection index of the artery, respectively. During my talk, I will give a live demo of the
devices and the software. If the audiences are interested in their artery condition, welcome to take a measurement!