Photoacoustic imaging reconstruction using combined nonlocal patch and total-variation regularization for...

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Computational Imaging

Background:

• Straight-line scanning in photoacoustic imaging (PAI) often results in significant artifacts due to missing data.
• Conventional total variation (TV)-based algorithms struggle with PAI reconstruction, leading to blurred images and loss of geometric detail.

Purpose of the Study:

• To develop an improved algorithm for photoacoustic imaging reconstruction that addresses artifacts from straight-line scanning.
• To enhance the quality and clarity of reconstructed PAI images in practical applications.

Main Methods:

• A combined nonlocal patch and TV-based regularization model (patch-TV) was developed for PAI reconstruction.
• A modified adaptive nonlocal weight function was employed for accurate patch similarity estimation and adaptive neighborhood searching.
• The optimization problem was integrated into a standard iterative PAI reconstruction framework.

Main Results:

• The patch-TV algorithm demonstrated significant qualitative and quantitative improvements over traditional TV-based and patch-based regularization algorithms.
• The proposed method showed enhanced peak signal-to-noise ratio and superior noise robustness.
• The algorithm exhibited efficient convergence and calculation speeds in simulations and experiments.

Conclusions:

• The developed patch-TV algorithm effectively mitigates artifacts in PAI reconstruction from straight-line scanning.
• The algorithm offers superior performance in terms of image quality, noise handling, and computational efficiency.
• Patch-TV is highly effective for practical photoacoustic imaging applications.