A Chambolle-Pock algorithm for sparse-view CT reconstruction via total generalized variation regularization and...

Area of Science:

• Medical Imaging
• Computational Imaging
• Image Reconstruction Algorithms

Background:

• Total Variation (TV) regularization causes staircase artifacts in CT images.
• Total Generalized Variation (TGV) regularization reduces staircase artifacts.
• Penalized Weighted Least-Squares (PWLS) enhances reconstruction accuracy.
• Traditional TGV-PWLS models have high computational complexity.

Purpose of the Study:

• To develop an efficient computed tomography (CT) image reconstruction algorithm for sparse-view data.
• To address the computational complexity of traditional TGV-PWLS models.
• To maintain and improve image reconstruction accuracy and detail preservation.

Main Methods:

• Reformulated the TGV-PWLS model as a saddle-point problem using dual variables.
• Transformed the saddle-point problem into an equivalent form.
• Solved the transformed problem using the Chambolle-Pock (CP) algorithm for a single-loop iterative solution (CP-TGV-PWLS).

Main Results:

• The CP-TGV-PWLS method significantly reduces reconstruction time compared to traditional algorithms.
• Demonstrated superior performance in detail preservation and structural feature reconstruction.
• Achieved higher reconstruction accuracy on phantom and clinical datasets.

Conclusions:

• The proposed CP-TGV-PWLS method offers an efficient and accurate solution for sparse-view CT image reconstruction.
• This single-loop algorithm effectively balances computational speed with image quality.
• The method shows promise for clinical applications requiring fast and precise CT imaging.