Iterative metal artifact reduction for x-ray computed tomography using unmatched projector/backprojector pairs

Area of Science:

• Medical Imaging
• Image Reconstruction
• Computational Imaging

Background:

• Metal artifacts significantly degrade X-ray computed tomography (CT) image quality, posing challenges for diagnosis.
• Conventional iterative reconstruction methods can introduce new artifacts around metal implants.
• Effective metal artifact reduction (MAR) is crucial for accurate image interpretation in CT examinations.

Purpose of the Study:

• To develop an iterative MAR method that suppresses newly introduced artifacts.
• To enhance image quality around metal implants in CT scans.
• To improve detail recovery in the presence of metallic objects.

Main Methods:

• An iterative strategy using unmatched projector/backprojector pairs was employed.
• A ramp filter was incorporated into the back-projection to address low-frequency inconsistencies.
• Constrained total variation (TV) minimization was integrated for enhanced efficiency, implemented as iFBP-TV and TV-FADM algorithms.

Main Results:

• The proposed methods effectively reduced streak metal artifacts in both simulated and real CT datasets.
• New artifacts near metal implants were successfully avoided.
• Quantitative metrics (RMSE, NMAD, UQI) and qualitative assessments showed superior performance of iFBP-TV and TV-FADM over conventional methods.

Conclusions:

• The developed iterative MAR algorithms outperform classical methods in streak artifact suppression.
• Edge structural information is preserved effectively, leading to improved image quality.
• Structures near metal implants are recovered due to the reduction of inconsistency-induced artifacts.