Related Experiment Video
Updated: Dec 29, 2025
Author Spotlight: Advancing CBCT and Digital Dental Image Integration with AI-Assisted Digitization
Published on: February 23, 2024
Creating Artificial Images for Radiology Applications Using Generative Adversarial Networks (GANs) - A Systematic
Vera Sorin1, Yiftach Barash1, Eli Konen1
1Department of Diagnostic Imaging, Chaim Sheba Medical Center, Affiliated to the Sackler School of Medicine, Tel-Aviv University, Emek Haela St. 1, Ramat Gan, Israel 52621; Deep Vision Lab, Sheba Medical Center, Tel Hashomer, Israel.
Generative adversarial networks (GANs) are revolutionizing radiology by creating realistic images for various applications. This review highlights their impact on image reconstruction, denoising, and data augmentation, improving clinical care and research.
Area of Science:
- Artificial Intelligence in Medical Imaging
- Deep Learning Applications
- Computer Vision in Radiology
Background:
- Generative Adversarial Networks (GANs) are advanced deep learning models.
- GANs excel at generating highly realistic synthetic images.
- These models have significantly impacted the field of computer vision.
Purpose of the Study:
- To conduct a systematic literature review on the applications of GANs in radiology.
- To identify and categorize the diverse uses of GANs within radiological imaging.
Main Methods:
- Systematic review adhering to PRISMA guidelines.
- Comprehensive search of electronic databases for studies on GANs in radiology.
- Inclusion of studies published up to September 2019.
Main Results:
- Analysis of 33 studies published between 2017 and 2019.
- Dominant applications include CT (18 studies) and MRI (10 studies).
- Key uses: image reconstruction/denoising (14 studies), data augmentation (6), modality transfer (8), and segmentation (5).
- All reviewed studies reported performance improvements using GAN-generated images.
Conclusions:
- GANs show increasing utility and study in diverse radiology applications.
- The technology facilitates novel data creation, enhancing clinical care, education, and research.
- GANs offer a promising avenue for advancing radiological practices.
