Solving the MPI reconstruction problem with automatically tuned regularization parameters

Area of Science:

• Medical Imaging
• Tomographic Techniques
• Image Reconstruction

Background:

• Magnetic Particle Imaging (MPI) is a non-ionizing tomographic technique offering high spatial and temporal resolution.
• MPI image reconstruction relies on iterative solvers and system matrices, often requiring manual regularization parameter tuning.
• Current manual regularization methods are time-consuming and depend heavily on user expertise and specific measurement conditions.

Purpose of the Study:

• To develop an automated method for MPI image reconstruction.
• To simplify the regularization process by reducing multiple parameters to a single, automatically determined parameter.
• To improve the efficiency and consistency of MPI image reconstruction.

Main Methods:

• Investigated the reduction of regularization parameters in MPI reconstruction to a single parameter.
• Developed and proposed a novel method for automatic regularization parameter selection.
• Validated the proposed method on multiple Magnetic Particle Imaging datasets.

Main Results:

• The proposed method successfully enables automatic reconstruction in MPI.
• Qualitative and quantitative validation demonstrated promising results across various MPI datasets.
• The automated approach addresses the challenge of varying regularization needs in MPI.

Conclusions:

• The developed method offers an efficient and effective solution for automatic MPI image reconstruction.
• Simplifying regularization to a single parameter significantly reduces manual effort and expertise required.
• This advancement holds potential for broader adoption and application of MPI in medical imaging.