Lowering the Computational Load in High Frame Rate 2-D Vector Flow Imaging

The market of portable, mid-range ultrasound imaging systems for ambulatory care has steadily grown. Hence, high-resource-demanding modalities must be adapted to be compliant with these systems. A high frame rate (HFR) 2-D vector flow imaging (VFI) algorithm was recently shown suitable for real-time implementation in a high-end system over a limited-size region of interest (ROI). In this work, we propose a solution to lower the computational load of HFR VFI without reducing the size of the ROI. The VFI algorithm estimates the displacement between kernels in consecutive frames. The pulse repetition interval (PRI) impacts the aliasing limit, but the needed estimation rate depends on the flow temporal gradients and may correspond to several PRIs. The VFI algorithm can thus be limited to pairs of consecutive frames, having a time-lag longer than the PRI. The higher the time-lag, the fewer the required estimates per second. The novel algorithm was tested in simulation on steady parabolic flow conditions. For time lags between 0.4 and 4 ms, the accuracy was in the range [2.4%, 3.1%] while the standard deviation slightly increased from 7.4% to 10.9%. On the other hand, the computational load decreases up to 85% for a time-lag equal to 10 times the PRI. Preliminary tests on pulsatile flow confirm the efficacy of such an approach.