Experimental evaluation of computerised tomography point spread function variability within the field of view:...

Area of Science:

• Medical Imaging
• Computational Physics

Background:

• Accurate characterization of the point spread function (PSF) is crucial for image quality assessment in computerised tomography (CT).
• Existing models may not fully capture the complexities of PSF variations within the scanner's field of view.

Purpose of the Study:

• To validate non-linear parametric models for CT PSF.
• To investigate the influence of model parameters and imaging conditions on PSF estimation.
• To experimentally assess PSF shape variations across the CT scanner's field of view.

Main Methods:

• Development and validation of two non-linear, 2D parametric PSF models: Gaussian (for positive PSF values) and damped cosine (for negative values).
• Fitting models to point source images to assess variance and modulation transfer function (MTF) errors.
• Quantification of blurring characteristics using shape and position parameters.

Main Results:

• The proposed models accounted for over 99% of the PSF signal variance.
• MTF errors were limited to 5% with appropriate model selection.
• Shape and position parameters effectively quantified blurring and achieved sub-pixel PSF localization.
• Anisotropy of the PSF was observed with an off-center point source.

Conclusions:

• Non-linear parametric models provide accurate and robust validation of CT PSF.
• These models effectively characterize PSF variations and anisotropy within the field of view.
• The validated models enhance the understanding and quantification of image reconstruction filter effects.