New beam hardened data correction and its application to artifact reduction in CT images

Area of Science:

• Medical Imaging
• Physics
• Computational Science

Background:

• Beam hardening in polychromatic CT systems causes artifacts due to energy-dependent photon absorption.
• These artifacts degrade image quality and diagnostic accuracy by violating the monochromatic assumption in filtered back-projection (FBP).

Purpose of the Study:

• To quantitatively estimate the mean energy of beam-hardened CT projection data.
• To develop a method using this mean energy to convert energy-dependent projections into corrected data.
• To verify the effectiveness of the mean energy-based correction method via simulations.

Main Methods:

• Derived a polynomial equation to mathematically determine the mean energy.
• Formulated the CT system using the Beer-Lambert law and an energy-integrated model.
• Applied integral theorems and power series expansions to obtain the polynomial equation for mean energy.

Main Results:

• Developed a novel correction method transforming energy-dependent data into monochromatic data using computed mean energy.
• Demonstrated significant reduction in beam-hardening artifacts in reconstructed CT images through simulations.
• Confirmed method's accuracy with complex materials (metals) and robustness across varying X-ray spectra, with minimal computational cost.

Conclusions:

• Presented a physically modeled, analytically approximated method for estimating mean energy to mitigate beam-hardening artifacts.
• The mean energy-based correction is a practical, computationally efficient solution for enhancing CT image quality.
• Further validation with experimental and clinical data is needed to confirm practical applicability.