Feasible images and practical stopping rules for iterative algorithms in emission tomography

Area of Science:

• Medical Imaging
• Image Reconstruction
• Computational Science

Background:

• Iterative tomographic image reconstruction using the maximum-likelihood estimator (MLE) can lead to image deterioration.
• Previous work introduced a stopping rule to mitigate this issue, but further investigation is needed.
• Understanding the causes of deterioration is crucial for accurate medical imaging.

Purpose of the Study:

• To introduce and define the concept of a "feasible image" in tomographic reconstruction.
• To examine the characteristics of the feasibility region in projection space.
• To develop and test a new rule for assessing the feasibility of reconstructed images.

Main Methods:

• Introduction of the "feasible image" concept based on the Poisson process of radioactive decay.
• Analysis of the shape and properties of the feasibility region.
• Development of a new rule to test reconstructions from real data for feasibility.
• Application of the new rule to reconstructions of the Hoffman brain phantom.

Main Results:

• The concept of a feasible image provides a criterion for acceptable tomographic reconstructions.
• The study defines the boundaries of the feasibility region in projection space.
• A new feasibility test rule is presented and demonstrated with real data.
• Comparative analysis of existing methods for addressing MLE image deterioration is provided.

Conclusions:

• The "feasible image" concept offers a principled approach to managing image deterioration in MLE tomographic reconstruction.
• The new feasibility rule enables objective assessment of reconstruction quality.
• This work contributes to improving the accuracy and reliability of medical imaging techniques.