Method for segmentation of cardiac signals based on four parameter sine fitting

This paper introduces a heartbeat segmentation method for cardiovascular signals based on 4-parameter sine fitting. The goal of the method is to segment unobtrusivelyacquired cardiovascular signals, without requiring an additional reference signal to synchronize the segmentation points and without strict assumptions about the subject´s heartbeat shape. Known methods rely on simultaneous electrocardiogram acquisition, or require an expert´s intervention, or assume strong premises regarding the signals´ shape and stability. The method proposed relies on a moving power-window combined with 4 parameter sine fitting to transform the cardiac signal and to obtain the fundamental frequency of oscillation, the cardiac rhythm. Validation tests were made using ballistocardiogram (BCG) and impedance cardiogram (ICG) signals from a diverse set of subjects. The validation signals were acquired imperceptibly to the subjects, freely seated on a wheelchair. The outcome of the method´s application on the validation signals confirmed its reliability.