key: cord-0852963-29gk2tvx
authors: Vinod, Dasari Naga; Prabaharan, S.R.S.
title: Data Science and the role of Artificial Intelligence in achieving the fast diagnosis of Covid-19
date: 2020-07-30
journal: Chaos Solitons Fractals
DOI: 10.1016/j.chaos.2020.110182
sha: cef6770691150d072c972d2b8e6b0e718dde6cc9
doc_id: 852963
cord_uid: 29gk2tvx

The rapid spread of novel corona virus (namely Covid-19) worldwide has alarmed a pandemic since its outbreak in the city of Wuhan, China in December 2019. While the world still tries to wrap its head around as to how to contain the rapid spread of the novel coronavirus, the pandemic has already claimed several thousand lives through out world. Yet, the diagnosis of virus spread in human has proven complexity. A blend of computed tomography imaging, entire genome sequencing, and electron microscopy have been at first adapted to screen and distinguish SARS-CoV-2, the viral etiology of Covid-19. There are a less number of Covid-19 test kits accessible in hospitals because of the expanding cases every day. Accordingly, it is required to utensil a self-exposure framework as a fast substitute analysis to contain Covid-19 spreading among individuals considering world at large. In the present work, we have elaborated a prudent methodology that helps identify Covid-19 infected people among the normal individuals by utilizing CT scan and chest x-ray images using Artificial Intelligence (AI). The strategy works with a dataset of Covid-19 and normal chest x-ray images. The image diagnosis tool utilizing decision tree classifier for finding novel corona virus infected person. The percentage accuracy of an image is analyzed in terms of precision, recall score and F1 score. The outcome depends on the information accessible in the store of Kaggle and Open-I according to their approved chest X-ray and CT scan images. Interestingly, the test methodology demonstrates that the intended algorithm is robust, accurate and precise. Our technique accomplishes the exactness focused on the AI innovation which provides faster results during both training and inference.

2019 version of "Coronavirus" is often represented with "novel," as it is another strain in the group of infections, we have all watched already. According to WHO's classification, coronaviruses have a place with a huge family run from the normal cold to hazardous ailments [1] . These sicknesses can blemish both in human and animals. The strain that began spreading in Wuhan in late 2019, the capital of China's Hubei province, is distinguished from two distinctive coronaviruses, SARS and MERS. The vast majority falls debilitated with Covid-19 will encounter gentle to direct side effects and recoup without unique treatment [2] .

Let us see how it spreads. The infection that causes Covid-19 is principally transmitted through droplets created when an infected individual coughs, sneezes, or breathes out. These droplets are too substantial to even think about hanging noticeable all around, and rapidly fall on floors or surfaces. You can be infected by breathing in the infection in the event that you are inside closeness of somebody who has Covid-19, or by contacting a polluted surface and afterward your eyes, nose or mouth. Covid-19 influences various individuals in various manners. Most contaminated individuals will create gentle to direct side effects. Common symptoms are fever, tiredness, dry cough. Some people may experience aches and pains, nasal congestion, runny nose, sore throat, diarrhoea [3] .

During normal course, it takes 5-6 days from when somebody is contaminated with the infection for side effects to appear, anyway it can take as long as 14 to 21 days. Individuals with mild side effects who are in any case healthy should self-isolate. Look for clinical consideration on the off chance that you have a fever, a cough, and difficulty breathing.

The analysis of COVID-19 is as of now a troublesome errand on account of inaccessibility of analysis framework all over the place, which is causing fear. Due to the restricted accessibility of COVID-19 testing kits, we need to depend on different conclusion measures.

Since COVID-19 assaults the epithelial cells that line our respiratory tract, we can utilize X-ray and CT images to investigate the soundness of a patient's lungs. The clinical specialists every now and then utilizes X-ray and CT scan pictures to analyze pneumonia, lung aggravation, abscesses, as well as developed lymph hubs [4] . Furthermore, nearly in all medical clinics have X-beam imaging machines, it could be conceivable to utilize X-ray and CT to test for COVID-19 without the committed test packs. Moreover, such routine imaging techniques poses disadvantages that X-ray investigation requires a radiology master and takes critical time, which is valuable when individuals are wiped out around the world. Along these lines, building up a robotized investigation framework is important to spare clinical experts important time. ratio is plausibly less [6] .

In this paper, a classification method is proposed to predict Covid-19 positive infected patient by using chest x-ray and CT scan images. At initial stage the dataset images are resized for faster understanding the machine next step resizing images are converted into RGB format.

Due to the prediction of the images, datasets are splitted into both training and testing. Final step to apply decision tree classifier along with AI to predict whether the image is normal or COVID 19 positive. The below fig. 3 In this paper, a framework dependent on deep learning is produced for the recognizable of COVID 19 as a characterization task. In this examination, we arranged two sets of dataset of Chest X-ray images. The first dataset contains 9 number of Covid-19 positive and 9 number normal X-ray chest images. These datasets are gathered from Kaggle repository. The normal dataset contains 104 images and covid-19 chest X-ray image dataset contains 297 images.

The two datasets are analyzed independently in the proposed models. We utilize this dataset for profound component extraction dependent on deep learning strategy. The below flow chart shows the whole process of the proposed model. 

In this investigation, we inspected the presentation of arrangement models for recognizable covid-19 positive with the help of AI models. The test considers were executed utilizing the Anaconda-Jupyter notebook tool all applications were run on a PC. With the help of AI and deep learning strategy predicted both chest X-ray and CT scan images. The outcome depends on the information accessible in the Kaggle and Open-I according to their approved X-ray pictures and CT scan images. Fig.7 . CT scan images predicted Covid-19 Positive with AI

Given the importance of plausible test methodology to diagnose COVID19 patents, it is imperative to identify quick testing protocol to diagnose the affected patients. The substance of the proposed model concerning the COVID-19 depends on the information accessible in WHO, the chest X-ray pictures and CT scan images are utilized for simulation intentions are gathered from Kaggle repository. For discovery of novel coronavirus utilizing X-ray pictures and CT scan images dependent on fusion of deep learning strategies and AI. For decision tree classifier we classify the feature extractions of the images and splitting for training and testing. The recommended analysis model for identification of COVID-19 is accomplished to the accuracy of 93% recall score in CT scan images and 88% of precision score in Chest xray images. Furthermore, work is underway, to involve app features to a doctor for predictive analysis via internet of things (IoT).

Detection of coronavirus Disease ( COVID-19 ) based on Deep Features

CoronaTracker: World-wide Covid-19 outbreak data analysis and prediction

Propagation analysis and prediction of the COVID-19

Predicting turning point, duration and attack rate of COVID-19 outbreaks in major Western countries

Modeling and Predictions for COVID 19 Spread in India

The authors are indebted to SRM University for supporting the work during covi-19 lockdown. One of us (DNV) is grateful to SRM University for the award of Research assistantship.

☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.