Fast linear least-squares method for ultrasound attenuation and backscatter estimation

Area of Science:

• Medical Physics
• Biomedical Engineering
• Acoustics

Background:

• Ultrasonic attenuation and backscatter coefficients are crucial acoustic parameters in medical imaging.
• Accurate estimation of these parameters has significant clinical applications.

Purpose of the Study:

• To propose a novel linear least-squares method for estimating ultrasonic attenuation and backscatter coefficients.
• To evaluate the method's accuracy, computational efficiency, and sensitivity to parameters.

Main Methods:

• A linear least-squares method fitting an ultrasound backscattered signal model to pulse-echo measurements.
• Simultaneous fitting in both frequency and depth dimensions.
• Analysis of parameter sensitivity (window length, overlap, bandwidth) and comparison with dynamic programming.

Main Results:

• The proposed method accurately estimates attenuation and backscatter coefficients.
• Inclusion of depth information enhances attenuation estimation accuracy.
• The method demonstrates superior accuracy and reduced computation time compared to a benchmark dynamic programming approach.

Conclusions:

• The developed linear least-squares method provides an efficient and accurate means for estimating ultrasonic attenuation and backscatter coefficients.
• Its low computational cost and high accuracy suggest suitability for real-time clinical ultrasound systems.