Efficient non-negative constrained model-based inversion in optoacoustic tomography.

Area of Science:

• Medical Imaging
• Biomedical Optics
• Computational Science

Background:

• Optoacoustic tomography (OPT) image reconstruction is sensitive to parameters like detector number and forward model accuracy.
• Image ambiguities, such as non-physical negative values, arise from these parameters in model-based inversion.
• Non-negative constraints are crucial for accurate optical absorption representation.

Purpose of the Study:

• To investigate the impact of non-negative constraints on model-based optoacoustic inversion algorithms.
• To introduce and analyze a novel conjugate gradient-based algorithm with non-negative constraints.
• To evaluate the accuracy, artifacts, and computational performance of constrained versus unconstrained inversion methods.

Main Methods:

• Analysis of state-of-the-art non-negative constrained algorithms for optoacoustic tomography.
• Development of a new conjugate gradient method incorporating non-negative constraints.
• Experimental validation using phantoms and small animal models.

Main Results:

• The introduced conjugate gradient algorithm demonstrates higher computational performance compared to other non-negative constrained methods.
• Non-negative constraints lead to improved image quality and quantitativeness.
• Experimental results show enhanced accuracy and reduced artifacts with the constrained approach.

Conclusions:

• Non-negative constraints are validated as an effective strategy for enhancing optoacoustic image accuracy and quantitativeness.
• The developed conjugate gradient algorithm offers computational efficiency alongside accurate reconstruction.
• Constrained inversion methods provide superior results compared to unconstrained approaches in optoacoustic tomography.