A motion artifact reduction method for dental CT based on subpixel-resolution image registration of projection data

Area of Science:

• Medical Imaging
• Dental Technology
• Image Processing

Background:

• High-resolution 3D CT imaging is vital for digital dentistry.
• Head motion during CT scans significantly degrades image quality, compromising digital dentistry applications.
• Existing methods often require external motion tracking devices.

Purpose of the Study:

• To introduce a retrospective motion artifact reduction (MAR) method for dental CT scans.
• To eliminate the need for external motion tracking devices.
• To improve the spatial resolution and efficacy of 3D image-based digital dentistry.

Main Methods:

• Motion information was extracted from projection data, assuming dominant rigid-body motions.
• Cross-correlation between successive projection data determined displacement for each view.
• Displacement was decomposed into translational and rotational components for projection data correction prior to CT reconstruction.

Main Results:

• The MAR method successfully corrected projection data based on motion amplitude.
• Experimental verification using phantoms showed significant improvement in image quality.
• Structural Similarity Index (SSIM) improved to 99%, and Normalized Absolute Error (NAE) reduced by 80-90% compared to motion-less images.

Conclusions:

• The developed retrospective MAR method effectively reduces motion artifacts in dental CT scans.
• This technique enhances image quality without requiring external motion tracking systems.
• The findings support the improved efficacy of 3D digital dentistry applications.