Few-view single photon emission computed tomography (SPECT) reconstruction based on a blurred piecewise constant...

Area of Science:

• Medical Imaging
• Computational Science
• Image Reconstruction

Background:

• Few-view single photon emission computed tomography (SPECT) poses reconstruction challenges.
• Existing methods like MLEM can struggle with limited projection data, leading to artifacts.

Purpose of the Study:

• To propose and characterize a novel sparsity-exploiting algorithm for few-view SPECT reconstruction.
• To evaluate the algorithm's performance against traditional methods like MLEM.

Main Methods:

• Developed a sparsity-exploiting algorithm modeling objects as piecewise constant with blurring.
• Validated using projection data from simulated objects and Monte Carlo simulations of phantoms.
• Compared reconstructions with varying design parameters against MLEM.

Main Results:

• The algorithm successfully recovered objects in noiseless simulations.
• Demonstrated higher correlation with true phantoms and reduced streaking artifacts compared to MLEM.
• Showed potential for accuracy even when the object model doesn't perfectly match the assumed blurring.

Conclusions:

• The proposed sparsity-exploiting algorithm shows preliminary feasibility for few-view SPECT.
• It offers potential benefits in reducing artifacts and improving reconstruction accuracy with limited data.
• Further optimization of parameters is needed to mitigate introduced artifacts.