The Use of Artificial Intelligence in Computed Tomography Image Reconstruction: A Systematic Review

*Corresponding author: Theresa Lee* and Euclid Seeram

Full-Text PDF:

systematic review



Current image reconstruction techniques in computed tomography (CT) such as filtered back-projection (FBP) and iterative reconstruction (IR) have limited use in low-dose CT imaging due to poor image quality and reconstruction times not fit for clinical implementation. Hence, with the increasing need for radiation dose reductions in CT, the use of artificial intelligence (AI) in image reconstruction has been an area of growing interest.


The aim of this review is to examine the use of AI in CT image reconstruction and its effectiveness in enabling further dose reductions through improvements in image quality of low-dose CT images.


A review of the literature from 2016 to 2020 was conducted using the databases Scopus, Ovid MEDLINE, and PubMed. A subsequent search of several well-known journals was performed to obtain additional information. After careful assessment, articles were excluded if they were not obtainable from the databases or not available in English.


This review found that deep learning-based algorithms demonstrate promising results in improving the image quality of low-dose images through noise suppression, artefact reduction, and structure preservation in addition to optimising IR methods.


In conclusion, with the two AI-based CT systems currently in clinical use showing favourable benefits, it is expected that AI algorithms will continue to proliferate and enable significant dose reductions in CT imaging.


Computed tomography (CT); Artificial Intelligence (AI); Image reconstruction (IN); Machine learning (ML); Deep learning (DL); Dose reduction.


AI: Artificial intelligence; CT: Computed tomography; ML: Machine learning; DL: Deep learning; FBP: Filtered back-projection; IR: Iterative reconstruction; MBIR: Model-based iterative reconstruction; LDCT: Low-dose computed tomography; FDA: U.S Food and Drug Administration; ANN: Artificial neural network; DNN: Deep neural network; CNN: Convolutional neural network; CNR: Contrast-to-noise ratio; SNR: Signal-to-noise ratio.