Efficient implementation of a local tomography reconstruction algorithm.

Area of Science:

• Medical Imaging
• Computational Science
• Image Reconstruction

Background:

• Tomography reconstruction generates detailed internal images but faces challenges with large datasets and artifacts.
• Interior tomography reconstruction, focusing on specific subregions, offers targeted artifact reduction.
• Existing methods struggle with computational demands for large-scale, high-resolution reconstructions.

Purpose of the Study:

• To present an efficient and optimized interior tomography reconstruction method.
• To address the computational challenges posed by increasing data volumes in tomography.
• To improve the accuracy and speed of local tomography artifact reduction.

Main Methods:

• Reformulating the reconstruction problem to minimize the number of variables.
• Implementing computationally efficient operators for optimization algorithms.
• Utilizing a known subregion to guide iterative refinement of the reconstruction.

Main Results:

• The proposed method demonstrates significant optimization for handling large data volumes.
• Reconstruction of [Formula: see text] slices is achieved in tens of seconds.
• The method surpasses the performance of equivalent exact local tomography techniques in speed and efficiency.

Conclusions:

• The developed interior tomography reconstruction method offers a computationally efficient solution.
• This approach effectively reduces local tomography artifacts, enhancing image quality.
• The optimized implementation makes high-resolution tomography reconstruction feasible for large datasets.