Fast reconstruction of flat region in a super-short scan based on MD-FBP algorithm

Area of Science:

• Medical Imaging
• Computed Tomography
• Image Reconstruction

Background:

• Circular cone-beam computed tomography (CT) faces reconstruction efficiency challenges, particularly with the minimum data filtered-backprojection (MD-FBP) algorithm.
• The MD-FBP algorithm, while useful for super-short scan problems, suffers from computationally intensive backprojection.

Purpose of the Study:

• To enhance the reconstruction efficiency and parallel performance of the MD-FBP algorithm for flat regions in super-short scan CT.
• To develop an improved MD-FBP method addressing computational bottlenecks in backprojection.

Main Methods:

• Developed an improved MD-FBP algorithm incorporating an integral operation with a fixed integral interval during backprojection.
• Applied the method to image reconstruction of flat regions in circular cone-beam CT, specifically for super-short scan scenarios.

Main Results:

• The improved method demonstrates enhanced reconstruction efficiency and parallel performance compared to the original MD-FBP algorithm.
• Reconstruction uncertainty is negligible for flat regions with thickness less than 0.0349 R (scanning radius).
• Thicker regions can be reconstructed by appropriately increasing the scanning radius.

Conclusions:

• The proposed method offers a significant improvement in reconstruction efficiency for flat regions in super-short scan CT.
• Numerical simulations and real data experiments validate the correctness and advantages of the developed technique.