A Real-Time GPU-Based Framework for Curved Array Ultrafast Ultrasound Doppler Imaging

Doppler flow imaging is a hybrid imaging method with 2D and Doppler images. The traditional dual-type segment scanning sequence has limited real-time display frame rate and ensemble size. Ultrafast Doppler imaging improves sensitivity to low-velocity microfluidic by providing a higher frame rate and relieving the problem of limited ensemble size. However, the uniform-type continuous scanning sequence, which uses the same echo data to obtain the two-mode images separately, ensures a high frame rate but at the expense of 2D imaging quality. We propose a dual-type mixed continuous scanning sequence, where each combined frame consists of one or more BW mode lines and a complete Color mode frame, and both BW mode and Color mode emission conditions can be configured individually. The sequence maintains the continuous acquisition of micro-vessels for a long period while outputting high-resolution BW mode tissue images, and improves the sensitivity to micro-vessels. Meanwhile, this paper proposes a GPU dual-rate processing framework suitable for this sequence, which can synchronize the acquisition results at 750 Hz and 150 Hz frame rates, and provide both power Doppler images sensitive to low-speed blood flow and correct velocity Doppler images for high-speed blood flow. This GPU processing framework offers the possibility of observing different velocity blood flows in the same scene in clinical ultrasound diagnosis. To verify the performance of the proposed method, in vivo experiments were performed on the renal and liver of a healthy volunteer. The results demonstrate that the framework can enhance the visualization of micro-vessels and support real-time velocity analysis of low-velocity blood flow in the abdomen using the convex array.