Image recovery in computer tomography from partial fan-beam data by convex projections

Area of Science:

• Image reconstruction
• Computational imaging
• Applied mathematics

Background:

• Image recovery is crucial in various scientific fields.
• Incomplete data poses significant challenges in obtaining accurate reconstructions.
• Existing methods may struggle with specific data limitations.

Purpose of the Study:

• To develop an effective image recovery algorithm for incomplete data.
• To explore the application of the convex projections method (POCS) in image reconstruction.
• To incorporate prior knowledge constraints to improve reconstruction fidelity.

Main Methods:

• Utilized the convex projections method (POCS) for image recovery.
• Considered diverse incomplete-data geometries, including limited source travel and beam-blocking opacities.
• Integrated prior-knowledge constraints, such as image vector directivity, to guide the reconstruction process.
• Leveraged the Toeplitz structure of operators for practical algorithm implementation.

Main Results:

• Successfully applied POCS to reconstruct images from incomplete datasets.
• Demonstrated the effectiveness of prior-knowledge constraints in enhancing reconstruction quality.
• The proposed algorithm showed practical implementability through efficient operator structures.

Conclusions:

• The convex projections method (POCS) is a viable approach for image recovery with incomplete data.
• Prior-knowledge constraints significantly improve the accuracy and robustness of image reconstruction.
• The developed algorithm offers a practical solution for challenging imaging scenarios.