Cone beam X-ray scatter removal via image frequency modulation and filtering

Area of Science:

• Medical Imaging Physics
• Image Processing
• Radiological Sciences

Background:

• Patient scatter degrades image quality in cone beam (CB) imaging.
• Existing scatter removal methods often require physical modeling or scatter measurements.

Purpose of the Study:

• To present a novel, rapid method for patient scatter removal in CB projection images.
• To eliminate the need for scatter measurement, physical modeling, or assumptions about scatter distribution.

Main Methods:

• A modulator grid is used to differentially frequency modulate primary and scattered photons.
• Non-linear Fourier domain filtering is employed to attenuate scatter in the image spectrum.
• The method's validity is verified using analytical models and simulations.

Main Results:

• Simulations show near-complete removal of even large amounts of scatter.
• Recovered images are largely artifact-free with a 10% RMS error.
• The method is effective even with ideal modulators.

Conclusions:

• Theoretical validity of scatter removal via spatial frequency modulation is confirmed.
• A trade-off exists between scatter rejection and detector resolution (max 0.41 resolution).
• This resolution limitation is often acceptable in medical CB imaging due to scatter's impact on resolution.