Optimization-based scatter estimation using primary modulation for computed tomography.

Area of Science:

• Medical Imaging
• Computational Imaging
• Radiological Physics

Background:

• Image quality in computed tomography (CT) is degraded by scatter radiation.
• Scatter correction is crucial but challenging.
• Primary modulation offers simultaneous primary and scatter acquisition but faces underdetermined separation problems.

Purpose of the Study:

• To propose an optimization-based scatter estimation (OSE) algorithm for accurate scatter estimation and correction in CT.
• To address the challenge of separating primary and scatter signals in primary modulation methods.

Main Methods:

• Developed an OSE algorithm incorporating an objective function for scatter and primary separation.
• Integrated prior knowledge into the optimization framework: positivity of primary, local smoothness of primary and scatter, penumbra identification, and scatter-contaminated data insights.
• Utilized edge-preserving weighting to enhance estimation accuracy near object boundaries.

Main Results:

• OSE demonstrated improved scatter estimation accuracy, particularly near object boundaries.
• Simulations showed OSE outperformed existing primary modulation algorithms in CT number accuracy and noise reduction.
• Clinical testing on cone beam CT confirmed OSE's effectiveness in scatter correction, even with misaligned modulators.

Conclusions:

• The OSE algorithm enhances the robustness and accuracy of scatter estimation and correction.
• OSE shows promise for scatter correction across various X-ray imaging modalities, including radiography, cone beam CT, and fourth-generation CT.