Bayesian speckle tracking. Part I: an implementable perturbation to the likelihood function for ultrasound...

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Signal Processing

Background:

• Clinical ultrasound relies on accurate displacement estimation.
• The Cramer-Rao lower bound (CRLB) traditionally limits estimator accuracy.
• Existing likelihood functions are ineffective for diffuse scattering in clinical ultrasound.

Purpose of the Study:

• To develop a more discriminative likelihood function for ultrasound displacement estimation.
• To overcome limitations of the classic likelihood function in diffuse scattering scenarios.
• To improve the accuracy and precision of clinical ultrasound displacement measurements.

Main Methods:

• Proposed a novel, perturbed likelihood function for displacement estimation.
• Compared the proposed function against the classic likelihood function.
• Evaluated functions using posterior probability density functions (PDFs) with a noninformative prior.
• Simulations used bulk motion, a 6λ tracking kernel, and 30 dB SNR.

Main Results:

• The new likelihood function assigned significantly higher probability to true displacement (0.22 ± 0.16) compared to the classic function (0.070 ± 0.020).
• Improvements were observed for bulk motion, acoustic radiation force-induced motion, and compressive motion.
• The proposed method demonstrated effectiveness at signal-to-noise ratios (SNRs) above 10 dB and kernel lengths from 1.5 to 12λ.

Conclusions:

• The proposed likelihood function offers a significant improvement over the classic approach for ultrasound displacement estimation.
• This new method enhances accuracy in various motion types and under typical clinical ultrasound conditions.
• The findings suggest a potential for more precise clinical ultrasound diagnostics through improved displacement estimation.