Abstract

Info

Author(s):: ZHENG Xiaohan1; ZHU Yan1; YANG Yueqi1; LI Xiaoou2; ZHOU Zhiyong3; 1. School of Medical Instrument and Food Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China 2. College of Medical Instruments, Shanghai University of Medicine & Health Sciences, Shanghai 201318, China 3. School of Design and Art, Shanghai Dianji University, Shanghai 200240, China

Keywords:: Keywords: ballistocardiogram heart rate detection non-contact peak detection

Abstract:: Abstract: In view of the lack of anti-interference in existing heart rate detection based on ballistocardiograms, a robust heart rate detection model is proposed. After compensating the interfered fragments in the original signal, J-peak is taken as feature waveform, and J-wave features are amplified based on the filtering of J-peak frequency band and enhanced differential processing. Finally, an improved peak extraction algorithm is designed to locate the heartbeat and then obtain the estimation of heart rate. The experimental results show that the model used in the study has good adaptability to ballistocardiogram signals in different situations. The average coverage rate of beat by beat detection is 98.3%, and the average accuracy reached 98.4%. The heart rate output is in high consistency with gold standard, with an error controlled within ±5%. Moreover, the real-time operation is satisfactory. The proposed method can provide an appropriate heart rate detection scheme for clinical or family health care.