Efficient CT metal artifact reduction based on fractional-order curvature diffusion

Area of Science:

• Medical Imaging
• Computational Imaging
• Image Processing

Background:

• Metal artifacts significantly degrade X-ray computed tomography (CT) image quality.
• Existing artifact reduction methods often struggle with complex metal shapes and varying densities.
• Accurate reconstruction of projection data is crucial for diagnostic reliability in CT.

Purpose of the Study:

• To develop and evaluate a novel metal artifact reduction (MAR) method for X-ray CT.
• To leverage fractional-order curvature-driven diffusion for information recovery in projection data.
• To compare the proposed method against established MAR techniques.

Main Methods:

• A fractional-order curvature-driven diffusion model was applied to X-ray CT projection data.
• Metal-corrupted projection data was treated as a damaged image for information recovery.
• Numerical schemes for the diffusion model were analyzed and implemented.

Main Results:

• The proposed fractional-order diffusion method demonstrated superior performance in artifact reduction.
• Quantitative evaluation using peak signal-to-noise ratio (PSNR) showed significant improvements over interpolation methods.
• Simulation results confirmed the effectiveness of the novel MAR approach.

Conclusions:

• Fractional-order curvature-driven diffusion offers a promising approach for metal artifact reduction in X-ray CT.
• The method effectively recovers lost information, enhancing image quality and diagnostic potential.
• This technique provides a valuable alternative to conventional projection interpolation methods.