A Survey of COVID-19 Detection From Chest X-Rays Using Deep Learning Methods

A Survey of COVID-19 Detection From Chest X-Rays Using Deep Learning Methods

Bhargavinath Dornadula, S. Geetha, L. Jani Anbarasi, Seifedine Kadry
Copyright: © 2022 |Pages: 16
DOI: 10.4018/IJDWM.314155
Article PDF Download
Open access articles are freely available for download

Abstract

The coronavirus (COVID-19) outbreak has opened an alarming situation for the whole world and has been marked as one of the most severe and acute medical conditions in the last hundred years. Various medical imaging modalities including computer tomography (CT) and chest x-rays are employed for diagnosis. This paper presents an overview of the recently developed COVID-19 detection systems from chest x-ray images using deep learning approaches. This review explores and analyses the data sets, feature engineering techniques, image pre-processing methods, and experimental results of various works carried out in the literature. It also highlights the transfer learning techniques and different performance metrics used by researchers in this field. This information is helpful to point out the future research direction in the domain of automatic diagnosis of COVID-19 using deep learning techniques.
Article Preview
Top

Introduction

Corona Virus disease 2019 (COVID-19) is caused by Severe Acute Respiratory Syndrome Corona Virus 2 (SARS-CoV-2) is the COVID-19 pandemic. Wuhan the city of China, where the disease was first detected in December 2019, led to a significant effect on human lives and their health. 188,620,082 total cases and 4,065,876 deaths worldwide were recorded by the mid of July 2021. India is now in second place after the USA, with maximum cases recorded being involved. To classify the virus in the human body, a special procedure called Real-Time Reverse Transcription–Polymerase Chain Reaction (RT- PCR) Test is used. In general, the COVID-19 virus collects in the throat or within a person's nose. The diagnostic method of the RT-PCR test begins with the collection of samples by swab from the specified parts of the body. By treating them with various chemical solutions, the cells and nucleus are lysed. Lastly, the sample contains deoxyribonucleic acid (DNA) and RNA only. It is a mixture of the individual's residual genetic material and the virus's RNA. Using a particular enzyme via reverse transcription, the RNA is then converted to DNA. To build a mixture, short fragments of DNA are also added. If the sample contains a virus, then the very less parts of the DNA match the output parts of the viral DNA. Then the hybrid input is put into an RT-PCR unit. By cycling through temperatures to ignite chemical reactions, the RT-PCR machine heats and cools the mixture. The target parts of viral DNA get their new identical copies through this process.

Repeated cycles are carried out with the sample by the RT-PCR COVID-19 Test Unit. This helps to copy the viral DNA target pieces. The number of copies of viral DNA would double at the end of each cycle. As a result, at the final stage of the scenario, approximately 35 billion new duplicates of the viral DNA parts from each variety of the virus are produced. Usually, this entire process of RT-PCR takes 3-4 days and if the person is suffering from COVID-19 the chances of spread from that person to multiple people are very high before one can check their result and as the cases are increasing rapidly lack of testing kits have become a major problem throughout the world. Taking this as a major problem, alternative methods are under research for finding out disease at a very early stage that is by using Chest X-ray to find out the virus in human’s lungs.

When using CT scans or X-rays images, the identification of COVID-19 symptoms in the lower portions of the lungs is more reliable than when using RT-PCR. Scans and chest X-ray tests may be replaced with RT-PCR tests in some situations. However, because of the comparatively lesser count of radiologists concerning current residents and the high amount of reappraisal of affected people who want to know the advancement of their disease, they cannot fix the issue exclusively. Increase in the pace of the procedure to resolve the difficulties of CT scans and X-rays and to support radiologists. This can be done through the design of advanced diagnostic systems using instruments of Artificial Intelligence (AI). The main goal is to bring down the time and effort needed to conduct COVID-19-positive patients' CT scans and X-rays and to assess the rate of progression of the disease. Radiological imaging is considered an effective screening technique for the diagnosis of COVID-19, and several authors have shown that the radiological history of COVID-19-related pneumonia is compatible with the clinical existence of the disease.

Figure 1.

Chest X-ray of a normal person, Pneumonia infected person, and COVID-19, the infected patient

IJDWM.314155.f01
Figure 2.

Class activation mapping of normal, pneumonia, and coronavirus patient

IJDWM.314155.f02

Complete Article List

Search this Journal:
Reset
Volume 20: 1 Issue (2024)
Volume 19: 6 Issues (2023)
Volume 18: 4 Issues (2022): 2 Released, 2 Forthcoming
Volume 17: 4 Issues (2021)
Volume 16: 4 Issues (2020)
Volume 15: 4 Issues (2019)
Volume 14: 4 Issues (2018)
Volume 13: 4 Issues (2017)
Volume 12: 4 Issues (2016)
Volume 11: 4 Issues (2015)
Volume 10: 4 Issues (2014)
Volume 9: 4 Issues (2013)
Volume 8: 4 Issues (2012)
Volume 7: 4 Issues (2011)
Volume 6: 4 Issues (2010)
Volume 5: 4 Issues (2009)
Volume 4: 4 Issues (2008)
Volume 3: 4 Issues (2007)
Volume 2: 4 Issues (2006)
Volume 1: 4 Issues (2005)
View Complete Journal Contents Listing