Truncation correction for oblique filtering lines.

Area of Science:

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

Background:

• Filtered backprojection (FBP) algorithms commonly use oblique filtering lines, which are susceptible to artifacts from limited scan fields of view.
• Data truncation along these oblique lines causes artifacts in reconstructed volumes, with a lack of existing correction methods for this specific scenario.
• Current FBP methods primarily address truncation along detector rows, leaving oblique filtering line truncation unaddressed.

Purpose of the Study:

• To introduce two novel methods for correcting data truncation artifacts caused by oblique filtering lines in FBP algorithms.
• To address the limitations of existing methods by providing solutions for oblique filtering line truncation.
• To enhance the quality of reconstructed volumes in CT imaging with limited scan fields of view.

Main Methods:

• Method 1 (basic approach): Employs a hybrid data extrapolation (water cylinder and Gaussian) in two preprocessing steps, independent of the reconstruction algorithm.
• Method 2 (kink approach): Integrates data extrapolation concepts into the reconstruction algorithm itself.
• Both methods utilize data extrapolation techniques to mitigate truncation effects.

Main Results:

• The proposed methods effectively handle data truncation occurring along oblique filtering lines.
• Experiments on simulated FORBILD head phantom data demonstrated the efficacy of the truncation correction techniques.
• The methods were tested using the theoretically exact M-line algorithm with a partial-circle-plus-arc trajectory.

Conclusions:

• The developed truncation correction methods successfully address artifacts arising from oblique filtering lines in FBP.
• These novel approaches are applicable to any FBP algorithm utilizing oblique filtering lines, not limited to theoretically exact algorithms.
• The presented techniques offer significant improvements for CT image reconstruction in scenarios with limited scan fields of view.