A short-scan reconstruction for cone-beam CT using shift-invariant FBP and equal weighting

Area of Science:

• Medical Imaging
• Computational Imaging
• Image Reconstruction

Background:

• Circular cone-beam (CB) CT is crucial for medical imaging.
• Short scan acquisition in CB CT can lead to artifacts.
• Existing reconstruction algorithms like FDK have limitations with short scans.

Purpose of the Study:

• To develop a novel 3D reconstruction formula for circular CB short scans.
• To suppress cone-beam artifacts effectively in limited-angle CT acquisitions.
• To provide an efficient and adaptable algorithm for clinical implementation.

Main Methods:

• Derived a 3D reconstruction formula using 1D shift-invariant filtering, CB backprojection, and equal weighting.
• Converted divergent projections to parallel projections for analysis via the central slice theorem.
• Incorporated a correction term with differential backprojection and 1D Hilbert transform to the FDK algorithm.

Main Results:

• The proposed formula reduces to the FDK algorithm for full scans.
• The correction term effectively suppresses CB artifacts in short scans.
• Demonstrated superior performance compared to the modified FDK algorithm with Parker's weighting through simulations and experiments.

Conclusions:

• The novel reconstruction formula offers improved artifact suppression for circular CB short scans.
• The algorithm is computationally efficient and easily adaptable from the FDK method.
• This advancement has the potential to enhance diagnostic accuracy in medical imaging using CB CT.