Minimum Variance Approaches to Ultrasound Pixel-Based Beamforming

Area of Science:

• Medical Imaging
• Signal Processing
• Ultrasound Technology

Background:

• Minimum Variance Distortionless Response (MVDR) beamforming is crucial for high-resolution imaging.
• Low echo signal-to-noise ratio (eSNR) and limited data samples degrade MVDR performance in pixel-based algorithms.
• Existing methods struggle with beamformer contributions far from the beam center.

Purpose of the Study:

• To integrate MVDR beamforming into a pixel-based algorithm for enhanced ultrasound imaging.
• To address the challenges of low eSNR and limited data samples in MVDR beamforming.
• To improve overall image quality in ultrasound applications.

Main Methods:

• Implemented a conventional minimum variance pixel-based (MVPB) beamformer performing MVDR post-superposition.
• Developed a novel coherent MVPB beamformer applying MVDR to individual transmits.
• Utilized pressure field analysis for improved data alignment and matrix estimation to overcome low-eSNR.

Main Results:

• The coherent MVPB beamformer significantly outperformed the conventional MVPB beamformer.
• Demonstrated superior performance in phantom and in vivo ultrasound studies.
• Showed comparable performance to state-of-the-art delay-and-sum algorithms in diffuse scattering regions.

Conclusions:

• The coherent MVPB beamformer effectively overcomes low-eSNR challenges in ultrasound imaging.
• This novel approach enhances image quality by improving beamformer contributions.
• The method shows promise for advanced ultrasound diagnostic applications, particularly in diffuse scattering environments.