Low-dose CT Image Restoration using generative adversarial networks

Kim, Jongchol; Kim, Jiyong; Han, Gyongwon; Rim, Choljun; Jo, Hyok

doi:10.1016/j.imu.2020.100468

Cited by 7 publications

(5 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…21 These algorithms make use of deep learning to transform raw sensor data to optimize the image. 22 In addition, GANs are also able to create synthetic data when there is a lack of data available to train AI algorithms. This data can then be used to train algorithms in the interpretive and noninterpretive domains.…”

Section: Noninterpretive Usesmentioning

confidence: 99%

“…GANs represent a subclass of neural networks that use two trained networks simultaneously, each serving different purposes—one directed toward image generation, the other toward discrimination 21. These algorithms make use of deep learning to transform raw sensor data to optimize the image 22. In addition, GANs are also able to create synthetic data when there is a lack of data available to train AI algorithms.…”

Section: Noninterpretive Usesmentioning

confidence: 99%

See 1 more Smart Citation

A Basic Primer of Artificial Intelligence for Radiologists

Stahl¹,

Blumer²

2022

Contemporary Diagnostic Radiology

View full text Add to dashboard Cite

Artificial intelligence (AI) comprises computer systems that behave in ways previously thought to require human intelligence. 1 AI and related technologies are increasingly prevalent in business and society and are beginning to be applied to health care. 2 Within health care, AI has increasingly influenced the field of radiology, and its role is likely only to grow in the future. Within radiology, AI has demonstrated benefits in the areas of image analysis and interpretation, various noninterpretive domains, and resident training. And yet, AI remains vaguely and incompletely understood by a great many practicing radiologists, radiology residents, and students considering a career in radiology. The purpose of this article is to describe the primary current and potential future applications of AI to the field of radiology and to define some of the key terms used in discussions of AI. This article is meant to provide readers with a clear, foundational

show abstract

Section: Noninterpretive Usesmentioning

confidence: 99%

Section: Noninterpretive Usesmentioning

confidence: 99%

A Basic Primer of Artificial Intelligence for Radiologists

Stahl¹,

Blumer²

2022

Contemporary Diagnostic Radiology

View full text Add to dashboard Cite

show abstract

“…In recent years, generative adversarial networks (GAN) have been extensively developed in the field of low-dose CT reconstruction [71,73,109,118,123,127,132,134,[146][147][148][149][150][151][152][153][154][155]. In contrast to convolutional neural networks (CNNs) in patches, [147] proposed denoising networks which are FCN-based using images in full size for training, and because they reused the underlying feature maps, the computational efficiency was very high.…”

Section: Other Applicationsmentioning

confidence: 99%

“…Then, the entire network became a type of generative adversarial network (GAN) with this complementary structure. Another current trend applied more complex loss functions so that observed smoothing artifacts can be overcome [18,54,71,95,100,109,123,129,130,134,146,150,151,154,[156][157][158][159]. Especially in [159], the loss function utilized in comparison has two pixel-level losses (mean square error and mean absolute error), the perceptual loss was based on the Visual Geometry Group network (VGG loss), and the one generated by Wasserstein for training gradient penalty adversarial network (WGAN-GP) adversarial loss, and their weighted sum.…”

Section: Other Applicationsmentioning

confidence: 99%

The use of deep learning methods in low-dose computed tomography image reconstruction: a systematic review

Zhang

Yang

Shi

2022

Complex Intell. Syst.

View full text Add to dashboard Cite

Conventional reconstruction techniques, such as filtered back projection (FBP) and iterative reconstruction (IR), which have been utilised widely in the image reconstruction process of computed tomography (CT) are not suitable in the case of low-dose CT applications, because of the unsatisfying quality of the reconstructed image and inefficient reconstruction time. Therefore, as the demand for CT radiation dose reduction continues to increase, the use of artificial intelligence (AI) in image reconstruction has become a trend that attracts more and more attention. This systematic review examined various deep learning methods to determine their characteristics, availability, intended use and expected outputs concerning low-dose CT image reconstruction. Utilising the methodology of Kitchenham and Charter, we performed a systematic search of the literature from 2016 to 2021 in Springer, Science Direct, arXiv, PubMed, ACM, IEEE, and Scopus. This review showed that algorithms using deep learning technology are superior to traditional IR methods in noise suppression, artifact reduction and structure preservation, in terms of improving the image quality of low-dose reconstructed images. In conclusion, we provided an overview of the use of deep learning approaches in low-dose CT image reconstruction together with their benefits, limitations, and opportunities for improvement.

show abstract

“…However, unlike previously described standard data augmentation approaches, which give minor changes, GANs may create realistic images that may offer additional variability to the training set. This ability of GANs [13] made it fascinating for researchers to work in medical imaging and has also been utilized in a diverse variety of applications [14] such as the denoising of low-dose CT images [15], synthesis of skin lesions [10], organ segmentation [16], and cross-modality transfer that includes MRI to CT [17] In this paper, we have investigated Pneumonia GAN (PGAN), a combination of an improved deep convolutional neural network (IDCNN) with a GAN for producing synthetic Xray images with the help of real datasets. This research aims to show the superiority of image augmentation based on GAN and discuss the model parameters along with the mathematical calculations are performed.…”

Section: Introductionmentioning

confidence: 99%

Data Augmentation using Generative Adversarial Network with Enhanced Deep Convolutional Neural Network for Pneumonia Chest X-ray Images

2024

Preprint

View full text Add to dashboard Cite

Pneumonia is a serious and life-threatening disease that results in difficulty breathing and needs to be diagnosed as soon as possible. Chest Xray is commonly used to diagnose pneumonia because they are cheap and readily available. Therefore the role of artificial intelligence in radiology has become a hot topic in using deep learning models. To train this model, a huge amount of datasets is required. To overcome this issue data augmentation technique has become a popular method for boosting the training data, particularly in medical images, where the data availability is insufficient. The generative adversarial network has become a popular method to generate synthetic data that looks like real data to increase the model’s performance compared to the traditional augmentation method. This paper focuses on a new GAN architecture i.e, Pneumonia GAN (a.k.a PGAN), to augment chest xray images using generative models. This novel PGAN architecture can successfully increase xray images. The performance of the novel PGAN model generates xray images that look like real images when compared with a conventional method. The comparison result proved that the proposed model generates an excess number of images compared to other techniques, and the mathematical calculation for each layer is discussed.

show abstract

Low-dose CT Image Restoration using generative adversarial networks

Cited by 7 publications

References 33 publications

A Basic Primer of Artificial Intelligence for Radiologists

A Basic Primer of Artificial Intelligence for Radiologists

The use of deep learning methods in low-dose computed tomography image reconstruction: a systematic review

Data Augmentation using Generative Adversarial Network with Enhanced Deep Convolutional Neural Network for Pneumonia Chest X-ray Images

Contact Info

Product

Resources

About