Image restoration for ring-array photoacoustic tomography system based on blind spatially rotational deconvolution

Area of Science:

• Biomedical imaging
• Image processing
• Optical physics

Background:

• Ring-array photoacoustic tomography (PAT) is vital for noninvasive biomedical imaging.
• PAT images often exhibit rotational blur and streak artifacts, limiting diagnostic accuracy.
• Existing methods struggle to fully address these image degradation issues.

Purpose of the Study:

• To develop an advanced image restoration technique for ring-array PAT.
• To mitigate spatially rotational blur and streak artifacts in reconstructed PAT images.
• To enhance the resolution and clarity of PAT images for improved biomedical applications.

Main Methods:

• Formulated image blur using spatially rotational convolution.
• Developed a regularized restoration model for blind spatially rotational deconvolution.
• Implemented an alternating minimization algorithm for image restoration and artifact removal.
• Proposed an image preprocessing method based on the deconvolution algorithm.

Main Results:

• The blind spatially rotational deconvolution algorithm significantly enhanced image resolution.
• The proposed method effectively removed streak artifacts from PAT images.
• Experimental results on phantoms and in vivo tissues validated the approach's efficacy.

Conclusions:

• The developed blind spatially rotational deconvolution method offers a significant improvement for ring-array PAT.
• This technique enhances image quality, paving the way for more accurate noninvasive biomedical imaging.
• The approach effectively addresses key limitations in current PAT image reconstruction.