Effect of errors in the system matrix on maximum a posteriori image reconstruction

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Computational Science

Background:

• Iterative image reconstruction is crucial for emission tomography (SPECT, PET).
• Accurate system matrices enhance image quality but are challenging to obtain precisely.
• Practical limitations often necessitate simplified system models.

Purpose of the Study:

• To analyze the impact of system matrix errors on maximum a posteriori (MAP) iterative image reconstruction.
• To derive a method for quantifying image artifacts caused by system matrix inaccuracies.
• To establish minimum system matrix accuracy requirements for emission tomography.

Main Methods:

• Developed an analytical expression using first-order Taylor series approximation.
• Analyzed errors in the system matrix for MAP iterative reconstruction.
• Validated theoretical findings with computer simulations.

Main Results:

• Derived an analytical expression for image artifacts due to system matrix errors.
• The expression helps determine the necessary system matrix accuracy.
• Computer simulations confirmed theoretical results in low-noise conditions.

Conclusions:

• System matrix accuracy is critical for minimizing artifacts in emission tomography.
• The derived analytical expression provides a quantitative basis for system matrix requirements.
• This work offers guidance for optimizing iterative reconstruction in SPECT and PET imaging.