Expediting model-based optoacoustic reconstructions with tomographic symmetries

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Computational Physics

Background:

• Accurate optoacoustic tomography (OAT) requires precise forward models for signal excitation, propagation, and detection.
• High-resolution OAT inversions often involve large matrices, leading to computationally intensive and time-consuming processes.
• Developing efficient, model-based approaches is crucial for advancing quantitative and dynamic OAT capabilities.

Purpose of the Study:

• To develop a method for simplifying and accelerating model-based inversions in optoacoustic tomography.
• To address the computational challenges associated with high-resolution image reconstruction in OAT.
• To enhance the quantitative and dynamic imaging performance of tomographic optoacoustic systems.

Main Methods:

• Introduced a method leveraging inherent symmetries in tomographic acquisition geometries for simplification and acceleration.
• Utilized polar image discretization of the time-domain optoacoustic forward model for cylindrical symmetries.
• Employed efficient storage and inversion strategies to optimize computational performance.

Main Results:

• Demonstrated fast and accurate model-based inversions in numerical simulations and experimental settings.
• Achieved a one-order-of-magnitude reduction in memory requirements for full-view detection schemes.
• Enabled high-resolution image reconstructions at video rates.

Conclusions:

• Exploited rotational symmetry in tomographic optoacoustic imaging systems to reduce computational demands.
• Facilitated the use of model-based algorithms with feasible computational requirements and rapid reconstruction times.
• Anticipated broad applicability and convenience for optoacoustic imaging systems exhibiting tomographic symmetries.