Monotonic penalized-likelihood image reconstruction for X-ray fluorescence computed tomography

Area of Science:

• Medical Physics
• Materials Science
• Imaging Science

Background:

• X-ray fluorescence computed tomography (XFCT) enables elemental distribution imaging.
• Accurate reconstruction requires correcting for photon attenuation.
• Unknown attenuation maps at fluorescence energies pose a significant challenge in XFCT.

Purpose of the Study:

• To develop a novel penalized-likelihood algorithm for XFCT image reconstruction.
• To address the challenge of unknown attenuation maps in XFCT.
• To improve the accuracy of elemental distribution imaging in XFCT.

Main Methods:

• A monotonic penalized-likelihood algorithm was derived.
• The algorithm iteratively updates element distribution and attenuation maps.
• A physically-motivated prior was incorporated to guide the reconstruction.

Main Results:

• The algorithm is guaranteed to increase the penalized likelihood at each iteration.
• The method allows for reconstruction with unknown fluorescence X-ray attenuation maps.
• The incorporated prior helps find a more accurate local maximum.

Conclusions:

• The developed algorithm offers a robust solution for XFCT image reconstruction with unknown attenuation.
• This advancement enhances the capability of XFCT for elemental analysis.
• The approach provides a pathway for more reliable quantitative imaging in XFCT.