Spatiotemporal Filtering for Synthetic Aperture Slow Flow Imaging

摘要

In this study, power Doppler images (PD) are obtained using a duplex synthetic aperture (SA) sequence with continuous data. Focused spherical waves are emitted in human in-vivo tissue with a handheld probe. To overcome the effects of motion blur, microvascular flow detection is enhanced by using a spatiotemporal higher order singular value decomposition (HOSVD) filter. Images of testicular parenchyma and vasculature in the carotid region were acquired from healthy volunteers. The data were acquired with an effective f_PRF of 2.08 kHz and beamformed using Synthetic Transmit Aperture beamforming. A second temporal axis was obtained by grouping the received signals in Doppler frames of 32 consecutive acquisitions. The time between Doppler frames is 15 ms (65 Hz), giving a sampling interval that represents the slow flow variation. Doppler frames were processed using a HOSVD filter, taking into account the 2 temporal dimensions and the spatial dimension. Power Doppler images were obtained for the soft tissue in proximity to the common carotid artery and the testicular parenchyma. In the latter case the imaged vessels are under 1 mm in diameter and the velocity of the flow within them is below 0.5 cm/s. This preliminary study with in-vivo data supports the use of sparse synthetic aperture sequences combined with eigen-filtering for slow flow imaging in human tissue.