A Synthetic Seismocardiogram and Electrocardiogram Generator Phantom

Abstract

The seismocardiogram (SCG) and electrocardio-gram (ECG) signals are two signals of cardiovascular origin containing important features for cardiac health assessment. Effective use of these signals requires recordings with acceptable signal to noise ratio. Studying the effects of external factors such as vibrations on these signals, and subsequent artifact removal algorithm design, remains a challenge due to lack of access to ground truth labels and human participant safety concerns. In this work, a synthetic SCG and ECG generator system is presented that enables data collection in environments that may be unsafe or inconvenient for human participants and offers ground truth labels along with the simulated recordings. The system was validated using real human SCG and ECG signals and showed >90%, and >98% input output correlations in both time and frequency domains for SCG and ECG signals respectively. Thus, the system is able to generate realistic SCG and ECG signals with clinically relevant amplitudes favorable for participant-free data collection in relevant environments.