Distributed three-dimensional simulation of B-mode ultrasound imaging using a first-order k-space method

Area of Science:

• Medical Imaging
• Computational Science
• Acoustics

Background:

• Realistic 3D ultrasound imaging simulations are crucial for research but computationally intensive.
• Serial computers struggle with the demands of complex, high-fidelity simulations.

Purpose of the Study:

• To develop and evaluate a parallel computing strategy for accelerating 3D ultrasound imaging simulations.
• To demonstrate the feasibility of performing complex 3D simulations using contemporary computing clusters.

Main Methods:

• A 3D k-space simulation method incorporating relaxation absorption and nonreflecting boundary conditions was employed.
• An allocation algorithm was developed for concurrent computation of scan lines across multiple nodes.
• An aperture-projection technique was introduced to reduce computational grid requirements for focused transducers.

Main Results:

• Parallel computing with 20 nodes reduced simulation time from 357.5 hours to 18.6 hours.
• The simulation accurately synthesized B-mode images, including realistic 3D refraction artifacts.
• Analysis confirmed the computational complexity and parallel efficiency for B-mode imaging.

Conclusions:

• The developed parallel computing approach makes fully 3D ultrasound imaging simulations practical.
• This advancement enables more efficient and detailed research in ultrasound imaging.
• Contemporary computing technology can support high-fidelity 3D ultrasound simulations.