Abstract
Velocity vector imaging (VVI) can suffer from poor signal-to-noise (SNR) conditions, eventually restricting the estimation of velocities at low SNRs. Cascaded dual-polarity wave (CDW) imaging has been shown to enhance blood-SNR, thereby improving cross-correlation-based VVI. However, the CDW decoding process is sensitive to motion, where imperfect summation and cancellation of pulses can lead to reduced amplitude gain and ghost pulses outside the main pulse. Multi/dual-angle transmits, combined with either autocorrelation- or cross-correlation-based axial velocity estimation, are commonly used for VVI. Both techniques have intrinsic limitations, such as speckle decorrelation and aliasing, and strengths, where autocorrelation is known to be more resistant to noise and cross-correlation results in more accurate estimates in high SNR conditions. This study evaluates the benefits of CDW imaging for both autocorrelation- and cross-correlation-based VVI using the FLUST simulation toolbox and experiments, including parabolic and rotational flow scenarios. The results indicate that autocorrelation performs consistently across the entire SNR range, while cross-correlation is more accurate and precise in high SNR conditions. CDW improves estimation performance in low SNR conditions, particularly for estimation of low velocities, with a more substantial performance boost for cross-correlation compared to autocorrelation. In general, CDW imaging shows to be beneficial for VVI, independent of the used velocity estimator.