Curve-Driven-Based Acoustic Inversion for Photoacoustic Tomography.

Area of Science:

• Biomedical Imaging
• Optical Physics
• Computational Imaging

Background:

• Quantitative photoacoustic tomography (PAT) relies on accurate model matrix computation for image reconstruction.
• Challenges in PAT include noisy signals and artifacts, hindering the establishment of effective model matrices.
• Existing methods like interpolated-matrix-model inversion (IMMI) can introduce interpolation-related errors.

Purpose of the Study:

• To introduce a novel curve-driven-based model-matrix inversion (CDMMI) method for calculating the model matrix in PAT.
• To evaluate the performance of CDMMI against conventional IMMI in various imaging scenarios.
• To demonstrate the capability of CDMMI in improving PAT imaging quality by mitigating noise and artifacts.

Main Methods:

• Developed the curve-driven-based model-matrix inversion (CDMMI) technique to compute the model matrix without interpolation.
• Compared CDMMI with the conventional interpolated-matrix-model inversion (IMMI) method.
• Validated the methods through numerical simulations, tissue-mimicking phantom studies, and in vivo small animal imaging.

Main Results:

• CDMMI demonstrated superior reconstruction accuracy compared to IMMI, especially as IMMI increased discrete points.
• The proposed CDMMI method effectively suppressed the negative influence of noise and artifacts.
• Improved overall imaging quality was observed in photoacoustic tomography using the CDMMI approach.

Conclusions:

• CDMMI offers a robust and accurate alternative for model matrix computation in PAT.
• The method eliminates interpolation errors, leading to enhanced reconstruction fidelity.
• CDMMI significantly benefits the quantitative accuracy and image quality of photoacoustic tomography.