SU-E-I-95: Multiscale Normalized Metal Artifact Reduction (MS-NMAR) in Computed Tomography.

Area of Science:

• Medical Imaging
• Computed Tomography (CT)
• Image Reconstruction

Background:

• Medical X-ray CT devices produce high-quality images, but metal artifacts degrade image quality.
• Conventional metal artifact reduction (MAR) methods often introduce new artifacts by not utilizing prior image information.
• Normalized MAR (NMAR) improves accuracy but is less effective for small-scale details.

Purpose of the Study:

• To introduce a novel multiscale normalized metal artifact reduction (MS-NMAR) method.
• To address the limitations of NMAR in handling small-scale details and reducing artifacts.
• To enhance the accuracy and quality of CT images in the presence of metal implants.

Main Methods:

• MS-NMAR generalizes NMAR using a multiscale framework, decomposing projection data and prior images into different scales.
• Normalization and linear interpolation are applied at each scale, with wider intervals for larger scales.
• Composite projection data are formed by summing components across all scales for image reconstruction.

Main Results:

• The efficiency of MS-NMAR was verified using real CT data.
• MS-NMAR demonstrated superior performance compared to conventional MAR and NMAR.
• Image quality was evaluated by inspecting regions around metal materials.

Conclusions:

• Both NMAR and MS-NMAR significantly outperform conventional MAR.
• MS-NMAR offers an advantage over NMAR by preventing the creation of new artifacts.
• The proposed MS-NMAR method effectively reduces metal artifacts while preserving image details.