Efficient 3-D Model-Based Reconstruction Scheme for Arbitrary Optoacoustic Acquisition Geometries.

Area of Science:

• Biomedical Imaging
• Optoacoustic Imaging
• Computational Imaging

Background:

• Optoacoustic tomographic sampling is limited by ultrasound sensor frequency-dependent directivity.
• Accurate 3-D modeling is crucial for compensating sensor directivity in optoacoustic imaging.
• Existing methods face computational and memory challenges with complex 3-D models.

Purpose of the Study:

• To introduce a 3-D model-based reconstruction method for optoacoustic imaging.
• To address the challenges of arbitrary detector shapes and sizes.
• To improve computational efficiency and reduce memory requirements.

Main Methods:

• Developed a 3-D model-based reconstruction algorithm for optoacoustic imaging.
• Implemented an efficient graphic processing unit (GPU)-based iterative inversion.
• Utilized on-the-fly model-matrix entry calculation with a look-up table to reduce memory usage.

Main Results:

• Validated superior imaging performance against standard methods using tissue-mimicking phantoms.
• Achieved significant improvements in spatial resolution and contrast-to-noise ratio.
• Demonstrated enhanced 3-D image quality in real tissue imaging of a human finger.

Conclusions:

• The proposed 3-D model-based method effectively compensates for sensor directivity in optoacoustic imaging.
• The GPU implementation offers a computationally feasible solution for complex optoacoustic systems.
• This approach significantly enhances image quality for biomedical applications.