Real-time, high-frame-rate, vector Doppler ultrasound imaging by a hybrid open platform

Ultrasound high-frame-rate (HFR) Vector Doppler Imaging (VDI) offers enhanced blood flow and tissue motion dynamics visualization. However, its clinical translation is hindered by a high computational load, which prevents real-time feedback. This work aims to implement and validate a novel heterogeneous open platform with a real-time HFR VDI technique based on two oriented coordinate systems. For each TX event, the scanner’s FPGAs perform parallel beamforming on the received echoes and reconstruct two raw images, one for each differently oriented coordinate system. Subsequently, the signals are demodulated and streamed to the onboard GPU module, which extracts the axial velocity components for both oriented directions through an HFR color flow mapping technique. Finally, the algorithm reconstructs the 2D vector field in a common Cartesian coordinate system by triangulating the information from the two oriented velocity maps. The proposed method was validated using a flow phantom with a 7 mm-diameter tube, where the maximum steady flow speed ranged between 4.4 cm/s and 8.4 cm/s. Results showed that the average bias and standard deviation on the estimation of the velocity module were smaller than 20% and 22%, respectively.