MUSiK: An Open Source Simulation Library for 3D Multi-View Ultrasound

Area of Science:

• Medical Imaging
• Computational Ultrasound
• Biomedical Engineering

Background:

• Diagnostic ultrasound is a vital, safe, and accessible medical imaging modality.
• Advanced applications like volume rendering and organ monitoring are hindered by hardware complexity and data scarcity.
• There is a critical need for tools facilitating rapid ultrasound prototyping and synthetic data generation.

Purpose of the Study:

• Introduce MUSiK, the first open-source library for multi-view acoustic ultrasound simulations of realistic anatomy.
• Provide a comprehensive toolset for the full image acquisition pipeline, from digital phantoms to image reconstruction.
• Demonstrate MUSiK's utility through diverse in vitro and in vivo use cases.

Main Methods:

• Development of MUSiK, an open-source ultrasound simulation library.
• Creation of digital anatomical phantoms and integration of diverse transducer types.
• Implementation of multi-view acoustic simulations and image reconstruction algorithms, including a Bayesian framework.

Main Results:

• Successful simulation of in vitro multi-view experiments, with analysis of image resolution and contrast.
• Demonstration of various in vivo imaging tasks: 2D kidney scans, 2D/3D echocardiography, 2.5D/3D tomography for lesion detection.
• Validation of a novel SNR-enhancing reconstruction algorithm within MUSiK's Bayesian framework.

Conclusions:

• MUSiK addresses the need for advanced ultrasound simulation tools.
• The library facilitates rapid prototyping, synthetic data generation, and exploration of novel imaging techniques.
• MUSiK is expected to accelerate technological development in the field of ultrasound imaging.