A PSF-Shape-Based Beamforming Strategy for Robust 2D Motion Estimation in Ultrafast Data

Abstract

This paper presents a framework for motion estimation in ultrafast ultrasound data. It describes a novel approach for determining the sampling grid for ultrafast data based on the system's point-spread-function (PSF). As a consequence, the cross-correlation functions (CCF) used in the speckle tracking (ST) algorithm will have circular-shaped peaks, which can be interpolated using a 2D interpolation method to estimate subsample displacements. Carotid artery wall motion and parabolic blood flow simulations together with rotating disk experiments using a Verasonics Vantage 256 are used for performance evaluation. Zero-degree plane wave data were acquired using an ATL L5-12 (f c = 9 MHz) transducer for a range of pulse repetition frequencies (PRFs), resulting in 0-600 µm inter-frame displacements. The proposed methodology was compared to data beamformed on a conventionally spaced grid, combined with the commonly used 1D parabolic interpolation. The PSF-shape-based beamforming grid combined with 2D cubic interpolation showed the most accurate and stable performance with respect to the full range of inter-frame displacements, both for the assessment of blood flow and vessel wall dynamics. The proposed methodology can be used as a protocolled way to beamform ultrafast data and obtain accurate estimates of tissue motion.

Investigation of the estimation accuracy of two-step block matching methods using envelope and RF signals for two-dimensional blood flow vector imaging

“…This paper focuses on the estimation accuracy of the block matching method based on a correlation function, [9][10][11][12][13][14][15][16] which is one of the most common techniques for 2D vector blood flow imaging. Several research groups have developed the multi-level block matching method.…”

Multicomponent Mechanical Characterization of Atherosclerotic Human Coronary Arteries: An Experimental and Computational Hybrid Approach

“…For highly accurate displacement estimation, the raw RF data instead of the demodulated RF data were used for the block-matching operation in this final iteration. Furthermore, the peak positions of the normalized cross-correlation functions were estimated on a subsample scale by 2D cubic interpolation of the cross-correlation peak ( Saris et al, 2018 ). No filtering was performed after the last iteration.…”

Carotid plaque stenosis, metabolism and flow dynamics: Important determinants of atherosclerotic risk?

“…As an alternative, novel ultrafast ultrasound imaging techniques, with up to 10.000 images/sec, could be used to estimate blood flow velocity vectors within the carotid artery. 6,7 The feasibility of this method has previously been shown in the carotid artery bifurcation of both healthy subjects and patients, with a proven reproducibility for the estimated velocity fields. 8 The further development of this technique might facilitate the implementation of flowbased decision making in clinical practice.…”