The Toast++ software suite for forward and inverse modeling in optical tomography

Area of Science:

• Biomedical optics
• Computational imaging
• Medical physics

Background:

• Diffuse optical tomography (DOT) is a non-invasive imaging technique.
• Accurate modeling of light propagation and robust inverse solvers are crucial for DOT.
• Existing software solutions may lack flexibility or advanced computational features.

Purpose of the Study:

• To introduce Toast++, an open-source software environment for diffuse optical tomography.
• To provide tools for simulating near-infrared light propagation and reconstructing optical properties.
• To facilitate rapid prototyping and development of DOT applications.

Main Methods:

• Finite-element solver for simulating light propagation in heterogeneous media.
• C++ implementation of forward solver with parallelization and GPU acceleration.
• Model-based iterative inverse solvers for absorption and scattering parameter reconstruction.
• Inclusion of regularization methods and prior knowledge incorporation.
• MATLAB and PYTHON bindings for user accessibility and script-based solutions.

Main Results:

• Toast++ enables accurate simulation of light propagation for various data acquisition systems (steady-state, time-, frequency-domain).
• The software provides robust inverse solvers for reconstructing optical properties from boundary measurements.
• High-performance computing features enhance solver efficiency.
• Flexible integration options (direct linking, MATLAB/PYTHON bindings) support diverse user needs.

Conclusions:

• Toast++ offers a comprehensive and efficient open-source platform for diffuse optical tomography research and development.
• The software's advanced features and flexibility accelerate the process of DOT application development.
• Toast++ empowers researchers with a powerful toolset for simulating, reconstructing, and visualizing DOT data.