Adaptive automated sinogram normalization for ring artifacts suppression in CT.

Area of Science:

• Medical Imaging
• Image Processing
• Computational Science

Background:

• Ring artifacts significantly degrade computed tomography (CT) image quality.
• These artifacts hinder automated post-processing tasks like segmentation and volume alignment.
• Existing artifact suppression methods are often manual, computationally intensive, or yield unsatisfactory results.

Purpose of the Study:

• To develop an automated, efficient, and computationally fast method for suppressing ring artifacts in CT projection data.
• To offer a preprocessing technique that balances effectiveness with speed and ease of use.
• To improve the precision of CT reconstructions by mitigating artifactual noise.

Main Methods:

• An automated projection data preprocessing method was developed, building upon the classical sinogram normalization technique.
• The method is designed to be adaptable to filtered sinograms.
• Key features include edge-preservation properties, validated on synthetic and real CT data.

Main Results:

• The proposed method achieved a 70.4% ring artifacts suppression percentage (RASP) on challenging open-access datasets, outperforming advanced techniques.
• Significant refinement in reconstruction quality was observed on real laboratory CT data.
• Performance was comparable or superior to existing manual ring artifact suppression methods.

Conclusions:

• The developed automated method provides an effective compromise between automaticity, efficiency, and computational speed for ring artifact suppression.
• The technique demonstrates adaptability and edge-preservation, leading to enhanced CT image reconstruction quality.
• This approach offers a valuable tool for improving the reliability and precision of CT imaging analysis.