Toward Real-Time GPU Implementation of Diverging Beam With Synthetic Aperture Technique With Non-linear Beamforming...

Area of Science:

• Medical Imaging
• Ultrasound Technology
• Computational Imaging

Background:

• Conventional ultrasound imaging has limitations in image quality outside the focal zone.
• Synthetic transmit aperture (STA) and non-linear beamforming improve image quality but require significant computational power.
• Real-time, affordable systems and STA imaging with curvilinear transducers are not yet widespread.

Purpose of the Study:

• To develop and evaluate a GPU-based, real-time, affordable system for curvilinear array ultrasound imaging.
• To implement and assess the diverging beam with synthetic transmit aperture technique (DBSAT) combined with non-linear beamforming.
• To explore the application of STA imaging with curvilinear transducers.

Main Methods:

• A GPU-based system was developed using an NVIDIA GeForce RTX 3060 GPU for beamforming.
• Experimental data were acquired using a tissue-mimicking phantom with DBSAT and conventional focused beamforming (CFB) sequences.
• Filtered delay multiply and sum (FDMAS) and delay-and-sum (DAS) algorithms were implemented for beamforming.

Main Results:

• DBSAT with FDMAS (DBSAT-FDMAS) using a curvilinear transducer improved image quality when the virtual source was closer to the transducer.
• Reducing the number of receive elements had minimal impact on image quality.
• DBSAT32-FDMAS demonstrated enhanced contrast-to-noise ratio (CNR) and generalized CNR (gCNR) compared to CFB-FDMAS with similar execution times.

Conclusions:

• The developed GPU-based DBSAT-FDMAS system offers enhanced ultrasound image quality with curvilinear transducers.
• This approach provides a viable solution for real-time, affordable advanced ultrasound imaging.
• Further optimization of virtual source positioning and receive element usage can be explored.