Model-based pulse pileup and charge sharing compensation for photon counting detectors: A simulation study.

Area of Science:

• Medical Physics
• Image Reconstruction
• Detector Science

Background:

• Pulse pileup (PP) and charge sharing (CS) are significant sources of spectral distortion in X-ray detectors.
• Accurate compensation for these effects is crucial for quantitative imaging, particularly in computed tomography (CT).
• Existing methods may introduce biases, especially under high count-rate conditions.

Purpose of the Study:

• To develop and evaluate a model-based algorithm (PCP) for compensating PP and CS effects.
• To assess the performance of the PCP algorithm against a comparative algorithm (LCP) using computer simulations.

Main Methods:

• Developed the PCP algorithm using cascaded models for CS and PP, maximizing Poisson log-likelihood with an exhaustive search.
• Developed a comparative LCP algorithm modeling loss of counts (LCs) and CS.
• Performed slab-based and CT-based simulations using a cadmium telluride detector, varying X-ray intensity and assessing bias and noise.

Main Results:

• PCP demonstrated minimal bias (<0.15 PCL) and noise (within 8% of CRB) in slab simulations, even at high probabilities of count loss (PCL up to 0.8).
• LCP exhibited significant biases (>±2 cm adipose) when PCL exceeded 0.15.
• CT simulations confirmed PCP's accuracy in reconstructing basis line integrals, density maps, and monoenergetic images, while LCP showed biases, especially in high-attenuation regions.

Conclusions:

• The PCP algorithm effectively compensates for PP and CS, providing statistically efficient and unbiased results for quantitative X-ray imaging.
• PCP is accurate across various count rates, outperforming LCP which introduces severe biases at high incident count rates (PCL ≥ 0.15).
• The developed PCP algorithm is suitable for applications requiring precise spectral compensation in CT and other X-ray imaging modalities.