From Linear System of Equations to Artificial Intelligence—The Evolution Journey of Computer Tomographic Image Reconstruction Algorithms

Document Type

Book Chapter

Publication Title

Indian Statistical Institute Series

Abstract

In 1970, the use of a linear system of equations in the form of algebraic reconstruction technique (ART) was an initial step to be used as an iterative reconstruction technique in the domain of computer tomography (CT). The method used by Gordon and Herman was popularly known as the Kaczmarz method in numerical linear algebra. There are many advantages of such an iterative reconstruction technique, however ART never took off for clinical trials due to a lack of computational power. Hence, the filtered back projection (FBP) method enjoyed as THE most popular and commercially used CT image reconstruction algorithm for decades. In 2009, the first iterative algorithm popularly known as the iterative algorithm in image space (IRIS) got FDA approval for commercial use as a CT image reconstruction algorithm. Such an advancement truly boosted researchers and commercial vendors to look for advanced iterative algorithms. As a result, the number of publications in the domain of iterative CT image reconstruction techniques is countless. Recent advancements in artificial intelligence have attracted researchers to explore the possibilities for the development of CT image reconstruction algorithms using the techniques of neural networks or deep learning. In this work, we intend our discussion to understand and explore the developments of major iterative reconstruction algorithms, especially as inverse problems, and optimization-based approaches for CT image reconstruction. We aim to understand what these algorithms have achieved in terms of radiation dose reduction while maintaining the image quality.

First Page

95

Last Page

115

DOI

10.1007/978-981-99-2310-6_5

Publication Date

1-1-2023

This document is currently not available here.

Share

COinS