Maximum-likelihood approach to strain imaging using ultrasound

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Acoustics

Background:

• Bioelasticity imaging systems require accurate strain estimation.
• Current methods face challenges in modeling complex ultrasonic waveform behavior.
• Understanding tissue deformation is crucial for diagnosing various medical conditions.

Purpose of the Study:

• To present a maximum-likelihood (ML) strategy for strain estimation.
• To establish a framework for designing and evaluating bioelasticity imaging systems.
• To develop a mathematical model for ultrasonic waveforms in strain imaging.

Main Methods:

• Combined continuum mechanics, signal analysis, and acoustic scattering principles.
• Developed a model incorporating 3-D object motion (affine transformations), Rayleigh scattering, and system response functions.
• Derived a likelihood function to express Fisher information matrix and variance bounds for displacement and strain estimation.
• Implemented the ML estimator as a generalized cross correlator using waveform warping and filtering.

Main Results:

• The ML estimator demonstrated performance approaching the Cramer-Rao error bound for small deformations.
• In soft tissuelike media experiments at 5 MHz and 1.2% compression, the predicted lower bound for displacement error was 4.4 microns.
• The measured standard deviation for displacement error was 5.7 microns, closely aligning with the theoretical bound.

Conclusions:

• The proposed ML strategy provides a robust framework for bioelasticity imaging.
• The developed mathematical model accurately represents ultrasonic waveforms for strain estimation.
• The ML estimator offers high accuracy for strain and displacement estimation in soft tissues.