Detection of premature ventricular beats from arterial blood pressure signal

Direct arterial blood pressure waveform is a rather complicated signal resulting from many factors that contribute to the waveform morphology. Among these factors, site of measurements, hemodynamics, cardiac output, stroke volume, and ventricular arrhythmias. Although direct and accurate measurements of systolic pressure, diastolic pressure, and mean arterial pressure can be done. However, accurate interpretation of the waveform is a very challenging problem. This paper presents a technique to detect premature ventricular beat (PVC) from the arterial pressure signal. The technique considers the arterial signal as an output of a linear time invariant system, then calculating the input impedance in frequency domain. The distribution of the real part of the impedance is used as the main feature that is fed into an unsupervised k-mean classifier. The technique showed promising results in terms of sensitivity (97.4%) and positive predictive value (96.2%) of the dataset used in the evaluation.