Comparison of time-domain displacement estimators for two-dimensional RF tracking

Area of Science:

• Medical Ultrasound
• Biomedical Engineering
• Signal Processing

Background:

• Multidimensional strain rate estimation in ultrasound imaging is crucial for assessing tissue mechanics.
• Accurate lateral strain rate estimation is challenging due to inherent noise in lateral velocity data compared to axial data.

Purpose of the Study:

• To identify the optimal estimator for tracking radiofrequency (RF) patterns in both axial and lateral directions for strain rate estimation.
• To compare the performance of various estimators in simulating 2-D velocity estimation.

Main Methods:

• Simulations were used to evaluate four estimators: cross-correlation, normalized cross-correlation, sum of absolute differences (SAD), and sum of squared differences (SSD).
• The study focused on two-dimensional (2-D) velocity estimation using a one-dimensional (1-D) kernel.

Main Results:

• Cross-correlation was not feasible for 2-D velocity estimation.
• Normalized cross-correlation, SAD, and SSD demonstrated accurate axial and lateral velocity estimation.
• Sum of squared differences (SSD) was identified as the preferred estimator for 2-D velocity estimation with smaller window lengths.

Conclusions:

• SSD offers superior performance for 2-D velocity estimation in ultrasound, particularly when using smaller tracking windows.
• This finding can lead to more robust and accurate lateral strain rate measurements in clinical applications.