A dual-domain regularization method for ring artifact removal of X-ray CT

Area of Science:

• Medical Imaging
• Detector Physics

Background:

• Photon-counting detectors (PCDs) offer enhanced spatial resolution and dose efficiency in CT imaging.
• Detector unit non-uniformity in PCDs is a significant challenge, leading to problematic ring artifacts in reconstructed CT images.

Purpose of the Study:

• To develop a novel dual-domain regularization model for effective ring artifact removal in PCD-based CT.
• To maintain the integrity and structural details of the original CT image during artifact correction.

Main Methods:

• A dual-domain regularization model is proposed, updating detector unit response inconsistency compensation coefficients using group and projection-view direction sparse constraints.
• Sparse constraints are applied in the image domain to further reduce ring artifacts.
• An alternating minimization method is employed to solve the proposed model, considering detector unit response and projection view relationships.

Main Results:

• The proposed method demonstrated superior performance in suppressing ring artifacts compared to existing techniques in both simulated and real PCD data.
• The method effectively preserved structural details within the CT images, outperforming conventional correction approaches.

Conclusions:

• The developed dual-domain regularization model effectively mitigates ring artifacts in PCD-reconstructed CT images.
• This approach enhances overall image quality and improves the reliability of clinical diagnoses derived from CT scans.