Systolic Phase Detection from Pulsed Doppler Ultrasound Signalusing EMD-DHT based Approach

Abstract

A new algorithm is developed to detect cardiac cycle’s systolic phase from the underwater construction workersDopplerultrasound signal. The method is on the basis of empirical mode decomposition (EMD), in combination with the discrete Hilbert transform (DHT). The EMD combined with DHT constitutes the Hilbert spectrum (HS) which is a fine-resolution time-frequency-energy representation of a nonstationary signal. Here EMD is used to calculate the intrinsic mode functions (IMFs), and thenthe DHT is employed to the IMFs to obtain three instantaneous components- frequency, amplitude and phase. Therefore, the HS is constructed from the normalized instantaneous frequencies and weighted sum of instantaneous amplitudes of the IMFs at the frequency bins. A new representation and interpretation of high graded signal is given to the time-frequency-energy distribution using HS. Finally systolic phase detection performance is evaluated in terms of sensitivity and positive predictivity.