Sub-Nyquist Sampling of Tissue Doppler Signals

Tissue Doppler ultrasound imaging (TDI) enables the quantification of cardiac function by estimating the velocity of the tissue from the phase shifts of the echoes. In order to estimate this velocity precisely, long consecutive pulse trains have to be transmitted in each direction, limiting TDI to a few measurements through the LV wall. This work aims to facilitate reduced time per velocity estimation and reduced sampling rate by transmitting non-uniformly spaced pulses in each direction and processing the signals within the compressed sensing (CS) framework. To this end a parametric model describing TDI signals is proposed. The proposed method was validated using realistic synthetic echocardiographic sequences. The correlation between the estimated velocities calculated from 50% of the pulses and those calculated from the entire dataset was 0.95. Spectral Doppler estimations were successfully performed with only 30% of the pulses and 40% of Nyquist frequency. Time gaps between pulses can be used for scanning in different directions in parallel.

Hence, the presented CS framework mitigates the tradeoff between spectral and spatial resolution.