Singular value decomposition-based 2D image reconstruction for computed tomography.

Area of Science:

• Medical Imaging
• Computational Science

Background:

• Inverse problems in imaging often lack analytical solutions, necessitating advanced reconstruction techniques.
• Traditional methods can be computationally intensive and slow, limiting their application for large datasets or real-time analysis.

Purpose of the Study:

• To develop and evaluate novel Singular Value Decomposition (SVD)-based 2D image reconstruction methods.
• To address limitations of existing methods by improving speed, accuracy, and memory efficiency for inverse problems.

Main Methods:

• Implementation of SVD-based algorithms for non-iterative 2D image reconstruction.
• Adoption of a multi-resolution strategy to manage memory constraints for large image reconstruction.
• Evaluation using modified Shepp-Logan, FORBILD head, and physical phantoms under various system configurations.

Main Results:

• SVD methods demonstrated high accuracy in reconstructing images from both standard and interior scan projections.
• The general SVD method showed superior performance compared to other SVD variants.
• Truncated SVD and Tikhonov regularized SVD effectively reconstructed regions-of-interest from internal scans.

Conclusions:

• SVD-based image reconstruction provides a fast, accurate, and flexible alternative to benchmark algorithms.
• The proposed methods are particularly advantageous for region-of-interest reconstruction tasks.
• The multi-resolution approach enhances the scalability of SVD methods for large-scale imaging applications.