2-D companding for noise reduction in strain imaging.

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Signal Processing

Background:

• Companding is a signal preprocessing technique used to improve correlation-based time delay measurements.
• In strain imaging, companding warps ultrasonic echo fields to enhance coherence and minimize decorrelation errors, a primary source of image noise.
• Current methods often use 1-D temporal stretching, a single-scale companding process.

Purpose of the Study:

• To introduce and evaluate a novel two-scale (global and local), 2-D companding technique for ultrasonic strain imaging.
• To assess the effectiveness of this 2-D companding method in improving target visibility and reducing noise.
• To compare the performance of 2-D local companding against traditional 1-D temporal stretching.

Main Methods:

• Developed a two-scale (global and local), 2-D companding technique.
• Utilized a sum-absolute-difference (SAD) algorithm for blood velocity estimation within the companding framework.
• Conducted experiments to demonstrate improvements in strain imaging using the proposed method.

Main Results:

• The 2-D local companding technique demonstrated significant improvements in target visibility in strain imaging.
• Displacement variance was reduced by two orders of magnitude when using 2-D local companding compared to 1-D temporal stretching.
• The technique effectively minimizes decorrelation errors, reducing noise in ultrasonic echo fields.

Conclusions:

• The proposed two-scale, 2-D companding technique offers a substantial advancement for ultrasonic strain imaging.
• This method effectively reduces noise and enhances image quality, particularly when tissue motion is confined to the scan plane.
• 2-D local companding provides superior performance over 1-D temporal stretching for reducing displacement variance.