Gaussian diffusion sinogram inpainting for X-ray CT metal artifact reduction

Area of Science:

• Medical Imaging
• Computational Imaging
• Image Reconstruction

Background:

• Implanted metal objects in patients cause severe streaking artifacts in X-ray computed tomography (CT) images.
• These artifacts degrade image quality, hindering accurate disease diagnosis.
• Effective metal artifact reduction (MAR) is crucial for clinical applications.

Purpose of the Study:

• To propose a novel Gaussian diffusion sinogram inpainting MAR algorithm for fan-beam CT.
• To leverage prior image information for artifact correction.
• To enhance diagnostic accuracy by improving image quality.

Main Methods:

• Developed a Gaussian diffusion sinogram inpainting algorithm incorporating prior image information.
• Utilized tissue-classified prior images for inpainting metal-corrupted projections.
• Employed gradient descent to solve the diffusion-based inpainting approach.
• Compared performance against interpolation and normalized MAR algorithms using simulated and clinical datasets.

Main Results:

• Subjective evaluation showed fewer secondary artifacts compared to conventional methods.
• Objective evaluation demonstrated the smallest normalized mean absolute deviation.
• The proposed method achieved the highest signal-to-noise ratio, indicating superior image quality.

Conclusions:

• The proposed Gaussian diffusion sinogram inpainting algorithm effectively reduces metal artifacts in fan-beam CT.
• The method shows significant improvements in image quality for both simulated and clinical datasets.
• This approach offers a promising solution for enhancing diagnostic capabilities in the presence of metallic implants.