Computed tomography ring artifact correction method with super-pixel segmentation and adaptive relative total...

Area of Science:

• Medical Imaging
• Image Processing
• Photon Counting CT (PCCT)

Background:

• Ring artifacts are a persistent problem in computed tomography (CT) imaging, often caused by hardware inconsistencies.
• These artifacts degrade image quality, impacting diagnostic accuracy and clinical utility.
• Photon counting CT (PCCT) presents unique challenges for artifact removal due to its advanced technology.

Purpose of the Study:

• To develop and validate a comprehensive method for removing ring artifacts in CT images.
• To specifically address the challenges of artifact correction in photon counting CT (PCCT).
• To preserve essential image details and grayscale balance during artifact removal.

Main Methods:

• A two-stage artifact correction approach combining super-pixel segmentation and adaptive Relative Total Variation (RTV).
• Super-pixel segmentation is used for high-intensity artifact removal.
• Adaptive RTV, optimized by mesh adaptive direct search, addresses low-intensity artifacts.

Main Results:

• Significant reduction in both high and low-intensity ring artifacts was observed.
• The method successfully maintained the structural integrity and grayscale balance of CT images.
• The proposed technique demonstrated versatility and robustness across various imaging scenarios and equipment.

Conclusions:

• The developed method effectively removes ring artifacts in CT images, particularly for PCCT.
• This advancement improves image quality and supports new applications of PCCT in precision medicine.
• The study provides publicly available code for reproducibility.