Sparse Convolutional Beamforming for 3-D Ultrafast Ultrasound Imaging

Area of Science:

• Medical Imaging
• Ultrasound Technology
• Signal Processing

Background:

• Real-time 3-D ultrasound (US) offers vital diagnostic capabilities but is hindered by high hardware costs and limitations in frame rate and image quality due to large data requirements.
• Existing 2-D sparse techniques show promise for element reduction and improved image quality.

Purpose of the Study:

• To introduce COBA-3D, a nonlinear beamformer for 3-D ultrasound imaging, building upon previous sparse convolutional beamforming algorithm (SCOBA) work.
• To develop SCOBA-3D, integrating 2-D sparse arrays for significant element reduction, enabling 3-D US with 2-D system resources.
• To present a method for designing 2-D fractal sparse arrays for efficient 3-D US system construction.

Main Methods:

• Developed COBA-3D, a nonlinear beamformer utilizing 2-D spatial convolution of in-phase and quadrature signals with diverging-wave transmission.
• Integrated 2-D sparse arrays into the COBA-3D framework to create SCOBA-3D.
• Designed 2-D fractal sparse arrays using a scalable and systematic approach.

Main Results:

• COBA-3D demonstrated improved image resolution and contrast compared to standard delay-and-sum beamforming.
• SCOBA-3D achieved significant element reduction, enabling high-frame-rate 3-D imaging with resources typical for 2-D setups.
• The proposed framework was validated using phantom and ex-vivo data, showing high-quality 3-D imaging capabilities.

Conclusions:

• The developed SCOBA-3D framework facilitates the creation of affordable, ultrafast 3-D ultrasound devices.
• This advancement has the potential to make high-quality 3-D US imaging more accessible in clinical settings worldwide.
• The systematic design of sparse arrays is key to overcoming hardware limitations in 3-D ultrasound.