Classification of the quality of wristband-based photoplethysmography signals
Authors:
- 1 May 2017
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 269-274
Abstract
Wearable technologies have made ubiquitous, non-invasive continuous monitoring of vital signs outside of the clinical setting possible. Convenient and user friendly embedded sensors in wearable technologies such as wristbands or watch, unlike cumbersome electrocardiogram Holter monitors, have made long term monitoring possible in normal home/home-care setting. However, such vital sign monitors are highly susceptible to motion artifacts and hence the quality of signal suffers. In order to develop a reliable automated technique to estimate vital signs, it is necessary to understand and estimate the quality of the acquired signal. In this paper, we compare the characteristics of signal corrupted by motion artifact against the artifact-free photoplethysmographic (PPG) signal acquired from an Empatica E4 wristband over 24 hours from 15 participants as the first step toward understanding and quantifying the quality of PPG signals. Using four 10 second segments of artifact-free and clean PPG signal from each participant, features that describe the signal are extracted. Bhattacharyya distance measure is used to rank the features that represent quality of the PPG signal. Using set of highly ranked features, a Naïve Bayes classifier is designed to quantify the quality of the PPG signal.Keywords
This publication has 5 references indexed in Scilit:
- Heart rate estimation using wrist-acquired photoplethysmography under different types of daily life motion artifact2015
- Dynamic time warping and machine learning for signal quality assessment of pulsatile signalsPhysiological Measurement, 2012
- Robust heart rate estimation from multiple asynchronous noisy sources using signal quality indices and a Kalman filterPhysiological Measurement, 2007
- Photoplethysmography and its application in clinical physiological measurementPhysiological Measurement, 2007
- Feature extraction based on the Bhattacharyya distancePattern Recognition, 2003