Photoacoustic image formation based on sparse regularization of minimum variance beamformer

Area of Science:

• Photoacoustic (PA) imaging
• Biomedical optics
• Signal processing

Background:

• Delay-and-sum (DAS) is standard for PA image formation but suffers from wide mainlobe and high sidelobes.
• Minimum variance (MV) beamforming offers improved resolution and contrast over DAS.
• Further enhancement is needed to suppress sidelobes more effectively in PA imaging.

Purpose of the Study:

• To introduce a novel Modified-Sparse-MV (MS-MV) algorithm for improved photoacoustic image reconstruction.
• To enhance sidelobe suppression and signal-to-noise ratio (SNR) in PA imaging.
• To provide an efficient and convex optimization solution for PA image formation.

Main Methods:

• Developed the Modified-Sparse-MV (MS-MV) algorithm by incorporating a ℓ1-norm constraint into the MV minimization problem.
• Interpreted the constraint as enforcing sparsity in the MV beamformed signals.
• Utilized a simple iterative algorithm to solve the resulting convex minimization problem.

Main Results:

• Numerical simulations demonstrated an average SNR improvement of 19.48 dB for MS-MV over MV.
• Experimental results with a wire-target phantom showed an SNR improvement of 2.64 dB for MS-MV compared to MV.
• The MS-MV algorithm effectively suppresses sidelobes, leading to clearer PA images.

Conclusions:

• The proposed MS-MV algorithm offers superior performance in photoacoustic image reconstruction compared to the conventional MV beamformer.
• MS-MV effectively reduces image artifacts and enhances image quality through improved sidelobe suppression.
• The algorithm provides a computationally efficient method for improving SNR in PA imaging.