Heart Rate Extraction of Ballistocardiogram Based on Hilbert-Huang Transformation

摘要

Introduction: Ballistocardiogram (BCG) was the signal that reflected the movement of human body caused by the mechanical activity of cardiovascular system, especially during the heart contraction. Compared to other methods to assess vascular healthy condition, acquisition of BCG didn't need any direct contact with human body. In this paper we tried to use Hilbert-Huang Transformation (HHT) to calculate the heart rate and detect the artifacts. Methods: HHT was a newly-developed method for nonlinear data analysis and Ensemble Empirical Mode Decomposition (EEMD) based HHT was a modified HHT method which used white noise to improve the analysis result. A device that could record BCG signal and ECG signal synchronously was built in our lab and 10 subjects' signals were collected and analyzed. EEMD based HHT was applied to BCG signal to calculate the heart rate. Heart rate calculated using ECG was used as standard value to verify the result calculated from BCG. Besides, BCG was easily affected by the body movement so we tried to use HHT to detect the artifacts in the BCG signal. Results: From our results, it showed that EEMD based HHT with a proper white noise level could be used to calculate heart rate in BCG. Artifacts in the decomposition component were enhanced in the decomposition components of EEMD and became easier to detect than that in original BCG signal. Conclusion: Therefore, HHT could help to calculate the heart rate, enhance and detect artifacts caused by movements of human body in BCG signal.