Heart rate variability (HRV) is the physiological variation of time between consecutive heart beats. Seismocardiography (SCG) is a non-invasive method of analyzing and recording cardiovascular vibrations transmitted to the chest wall. Mobile devices can monitor health parameters thanks to embedded sensors. In this study we compare HRV indices calculated from heart beats obtained from reference beats and tested beat detection algorithm on three SCG signals. Proposed algorithm consists of signal preprocessing, RMS envelope calculation and peak finding. Algorithm performance was compared to reference beat detector and defined as sensitivity (Se) and positive predictive value (PPV). We achieved average Se = 0.958 and PPV = 0.912, and in the best case Se =1.000 and PPV =1.000 for tested beat detector. HRV indices are influenced by the performance of heart beat detector. The smallest differences between HRV indices calculated from heart beats determined by different detectors were achieved for mean inter-beat interval, RMSSD, pNN50, PLF, PHF and LF/HF.
|Title of host publication||Advances in Intelligent Systems and Computing : Information Technology in Biomedicine. ITIB 2019.|
|Publication status||Published - 2019|
|Event||7th International Conference on Information Technology in Biomedicine|
- Kamień Śląski, Poland
Duration: 18.06.2019 → 20.06.2019
Conference number: 229749