Generation of Novel Synthetic Portable Chest X-Ray Images for Automatic COVID-19 Screening

Generation of Novel Synthetic Portable Chest X-Ray Images for Automatic COVID-19 Screening

Daniel Iglesias Morís, Joaquim de Moura, Jorge Novo, Marcos Ortega
Copyright: © 2022 |Pages: 34
DOI: 10.4018/978-1-6684-2304-2.ch008
OnDemand:
(Individual Chapters)
Available
$37.50
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

The diagnosis and the study of the evolution of COVID-19 is crucial to tackle the challenge that this disease represents for healthcare services. Chest x-ray imaging allows us to visualize the pulmonary regions, where COVID-19 causes its main affectation. In order to reduce the risk of cross-contamination, a crucial aspect in the pandemic, portable chest x-ray devices are advantageous being easier to decontaminate in comparison with the fixed machinery, despite offering a lower image quality. Furthermore, the recent emergence of COVID-19 implies a data scarcity that must be tackled. In this chapter, the authors present the analysis of a strategy that generates novel synthetic portable chest x-ray images using the CycleGAN, an architecture for image translation that is trained with unpaired data. The novel set of images is then added to the original dataset, improving the performance of the classification model.
Chapter Preview
Top

Introduction

The COVID-19 is a pathology that can affect several parts of the body, while its main affectation is located on the respiratory tissues. The pathogen that causes this disease is known as SARS-CoV-2, a type of coronavirus, which is highly contagious, reason why it was rapidly spread worldwide forcing the World Health Organization to declare the COVID-19 as a global pandemic in 12th March 2020 (Ciotti et al., 2020). Currently, more than 242 million confirmed cases alongside more than 4.9 million deaths were reported worldwide (Coronavirus Resource Center, Johns Hopkins, 2020). There are several techniques to diagnose the COVID-19 but the gold-standard is the RT-PCR test (Tahamtan & Ardebili, 2020). However, it could be interesting for clinicians to have more information than the simple diagnostic provided by this technique. The main affectation of the COVID-19 is located in the pulmonary area and, therefore, chest X-ray imaging can be very useful to visualize this region to diagnose the pathology, to study its severity and to understand its evolution (Jacobi et al., 2020), as it has been widely used in the last decades to diagnose other typical pulmonary diseases and detect pathological structures as is the case, for example, of the pneumonia (Fiszman et al., 2000), the tuberculosis (Van Cleeff et al., 2005), the fibrosis (Puderbach et al., 2007) or the lung nodules (Wei et al., 2002). In the context of the COVID-19, this image modality can be used as a complement to the diagnostic result of the RT-PCR. With all these ideas into account, we can conclude that chest X-ray emerges as a powerful approach to visualize the affection of this pulmonary disease. Despite its lower quality in comparison with other advanced methods, as the Computerized Tomography (CT) (Hayden & Wrenn, 2009), the chest X-ray captures are easier and cheaper to perform. On the other hand, as the SARS-CoV-2 is easily spread, it is very important to decontaminate the capture devices. In the same way, many patients are bedridden and, therefore, it is impossible for them to move to the radiology room. To solve these problematics, radiologists are recommended to use portable chest X-ray devices instead of the fixed machinery (Kooraki et al., 2020), as this kind of devices are easier to decontaminate and can be moved to where the patient is placed. However, in some cases, clinicians could decide to use more advanced techniques that provide a more detailed visualization of the pulmonary regions if it is necessary, in order to have a more precise localization of the pathological structures, despite the higher cost and difficulty of the captures. This is the case of the CT, that provide a 3-dimensional visualization of the captured region.

Key Terms in this Chapter

Deep Learning: Type of machine learning strategies that allows to create powerful classification and regression models able to deal directly with raw data.

Computer-Aided Diagnosis: Systems that are developed to improve the comprehension of medical images and help clinicians to make decisions.

Screening: Medical strategy that performs a diagnostic test on people that are considered healthy a priori with the aim to detect a possible disease on its mild stages.

Image Translation: Algorithmic process that converts an image from a source domain to another different target domain (for example, convert the image of an orange to an image of an apple).

Chest X-Ray: Medical image modality that allows to visualize the inside of the thoracic region of a patient using a low dose of ionic radiation.

Computerized Tomography: Medical image modality that combines a set of X-ray slices to get a 3-dimensional representation of a particular studied region of the body.

Portable X-Ray Devices: X-ray capturing devices that can moved to where the patients are placed.

Complete Chapter List

Search this Book:
Reset