An Effective Deep Learning Model for COVID-19 Detection from Chest X-Ray

International Journal of Research and Innovation in Social Science (IJRISS) | Volume V, Issue XI, November 2021 | ISSN 2454–6186

An Effective Deep Learning Model for COVID-19 Detection from Chest X-Ray

Simon Emmanuel Ikoojo¹, Deme Chuwang Abraham², Nentawe Gurumdimma^3
1,2,3Department of Computer Science, University of Jos, Plateau State, Nigeria

Abstract: A new viral disease that easily spreads was in December 2019 discovered in Wuhan city in China and was named by the World Health Organization (WHO) as COVID-19. The symptoms can be either mild or severe and mostly in older people who have hypertension, diabetes, and heart or lung disease. Early screening has proved to be effective in reducing the spread and the RT-PCR test is been employed for testing which is expensive and time-consuming. Deep learning using CNN on Chest X-rays can be used to detect the infection.
In this paper, three deep learning models (VGG16, Xception, and InceptionV3) were proposed for detecting COVID-19. These models were pretrained using images from ImageNet with the proposed Inception model achieving the highest accuracy of 98.28%. The f1-score for Xception, VGG, and Inception approaches are 98%, 95%, and 95% respectively. The proposed approaches achieved a precision score of 100%, 100%, and 96% in classifying COVID-19 cases for Inception, Xception, and VGG16 respectively.

Keywords: Deep learning, COVID-19, Chest X-Ray, RT-PCR test, CNN

I. INTRODUCTION

Anew viral disease that easily spreads was in December 2019 discovered in Wuhan city in China and was named by the World Health Organization (WHO) as COVID-19 [1]. Infection is transmitted mostly by droplets either saliva or mucous landing on the mouth, nose or eyes [2]. The symptoms can be either mild or severe and mostly in older people who have hypertension, diabetes, and heart or lung diseases [3]. The most common symptoms of this viral syndrome are fever, dry cough, fatigue, aches and pains, loss of taste/smell, and breathing problems[4]
According to the World Health Organization as of the 13th of October, 2021, there are about 238,229,951 confirmed cases of COVID-19, including 4,859,277 deaths [5]. Of the figures presented above, Nigeria accounts for about 208,153 confirmed cases with about 2,756 deaths [6]. As the COVID-19 epidemic has become a global pandemic, real-time analysis of epidemiological data is required to prepare society for better disease response plans [7].About 15% of COVID-19 cases are severe(Sharma & Sharma, 2020). This implies that Oxygen may be required for them to be treated in the hospital.

Read More