Autocorrelation-Guided Cross-Correlation for Ultrasound Shear Wave Elastography
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional motion tracking techniques in ultrasound shear wave elastography face challenges such as aliasing in autocorrelation and high computational burden in cross-correlation, leading to inaccurate displacement estimation due to phase aliasing and peak hopping, especially in clinical applications where patient motion can exceed the shear wave signal.
Innovation Solution
A two-step motion-tracking approach combining autocorrelation and cross-correlation, where autocorrelation is used to derive the initial phase and calculate axial displacement, followed by cross-correlation with a reduced search window to refine the displacement estimation, thereby overcoming aliasing and computational inefficiencies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If autocorrelation is used for motion tracking, then computational efficiency is improved, but displacement estimation accuracy deteriorates due to aliasing
Solution Approach 1:
The motion tracking process is segmented into two distinct stages: first, autocorrelation is used to obtain a coarse displacement estimate efficiently; second, cross-correlation is applied to refine this estimate and eliminate aliasing effects. This segmentation allows each method to operate in its optimal performance range.
Solution Approach 2:
Autocorrelation is performed as a preliminary step to generate an initial displacement estimate that serves as a starting point for the subsequent cross-correlation refinement. This preliminary action reduces the search space for the final accurate measurement.
2Measurement precision
If cross-correlation is used for motion tracking, then displacement estimation accuracy is improved, but computational burden increases
Solution Approach 1:
Cross-correlation is applied locally around the autocorrelation estimate rather than over the entire possible displacement range. This localized application maintains high accuracy while significantly reducing computational burden by limiting the search window.
Solution Approach 2:
The autocorrelation estimate serves as a preliminary result that guides the cross-correlation search, allowing the computationally intensive cross-correlation to focus only on a small region around the expected displacement, thereby improving efficiency.
3Reliability
If large search area is used in cross-correlation, then aliasing is avoided, but computational time increases
Solution Approach 1:
The search area problem is segmented by using autocorrelation to identify the likely displacement region first, then applying cross-correlation only within this restricted region. This avoids the need to search the entire large area while maintaining reliability.
Solution Approach 2:
Autocorrelation performs a preliminary search that identifies the approximate displacement location, allowing the subsequent cross-correlation to avoid searching unnecessary areas and thus reducing computational time while maintaining aliasing avoidance.
4Device complexity
If constant center frequency is assumed in Doppler-based approaches, then processing simplicity is improved, but displacement estimation accuracy deteriorates due to frequency dependent attenuation
Solution Approach 1:
The patent replaces Doppler-based frequency analysis (which requires center frequency assumptions) with autocorrelation-based phase analysis. This substitution eliminates the need for center frequency estimation and associated corrections, simplifying processing while improving accuracy.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This method provides more accurate and robust displacement estimation, reducing variance and bias, and significantly reducing computational time while avoiding the need for explicit estimation of mean frequency, thus improving the overall performance of shear wave elastography.
Implementation Method 1
Due to the effect of acoustic radiation force, the push pulse causes tissue in the focal area to move away from the probe surface, simultaneously establishing a shear wave propagating away from the focal region in a direction perpendicular to the push beam.
Implementation Method 2
the displacement of the structures of interest induces phase shift on successive high frequency ultrasound echoes backscattered by the moving medium
Implementation Method 3
Time-shift by cross-correlation estimates time delays by cross-correlating, from one pulse to another pulse using radiofrequency (RF) data or complex signals conveyed from RF data.
Implementation Method 4
As ultrasound waves propagate through soft tissue, the spectrum experiences a downshift due to the frequency dependent attenuation.
Data Source
AI summary
Ultrasound motion-estimation includes issuing multiple ultrasound pulses, spaced apart from each other in a propagation direction of a shear wave, to track axial motion caused by the wave. The wave has been induced by an axially-directed push. Based on the motion, autocorrelation is used to estimate an axial displacement. The estimate is used as a starting point (234) in a time-domain based motion tracking algorithm for modifying the estimate so as to yield a modified displacement. The modification can constitute an improvement upon the estimate. The issuing may correspondingly occur from a number of acoustic windows, multiple ultrasound imaging probes imaging respectively via the windows. The autocorrelation, and algorithm, operate specifically on the imaging acquired via the pulses used in tracking the motion caused by the wave that was induced by the push, the push being a single push. The algorithm may involve cross-correlation over a search area incrementally increased subject to an image matching criterion (S358).