5487
N. Sasirekha
a, G. Ravi
b, J. Harirajkumar
c, C. Kanmani
daAssociate Professor, Department of ECE, Sona College of Technology, Tamil Nadu, India. E-mail: [email protected]
bAssociate Professor, Department of ECE, Sona College of Technology, Tamil Nadu, India. E-mail: [email protected]
cAssociate Professor, Department of ECE, Sona College of Technology, Tamil Nadu, India. E-mail: [email protected]
dPG Scholar, Department of ECE, Sona College of Technology, Tamil Nadu, India. E-mail: [email protected]
Article History: Received: 11 January 2021; Revised: 12 February 2021; Accepted: 27 March 2021; Published
online: 10 May 2021
Abstract: Corona virus (COVID - 19) plays a crucial role in the biological disaster of human beings right now. In 2019,
corona virus had been identified in mammals of Wuhan, china and later it started to spread among the people in china and as a result it turned out as a pandemic worldwide. Corona virus causes various respiratory disorders in human beings and it produces severe consequences for the persons if affected along with comorbities. The infection in turn depends on the age of the person and the also on the immune system to fight against it. The viral infection in humans was confirmed through clinical Procedures. The PCR Polymerase Chain Reaction) gives us accurate results, but PCR technique is a time consuming process. As far as CT (Computed Tomography) scan has been used to study the infections in the internal organs of the infected people. In advancement to this AI (Artificial Intelligence) is used as tool to produce better results in the finding the infected areas compared with X-ray and CT scans. By employing AI, on the basis of internal lung lesions, the proposed system can identify, whether the infection is due to COVID – 19 or any other respiratory illness like lung cancer, pneumonia.
Keywords: COVID-19, CT Scan, Artificial Intelligent, PCR (Polymerase Chain Reaction).
1. Introduction
Corona virus also known as COVID-19, began to spread from the city wuhan, china in 2019. RNA (ribonucleic acid) viruses belonging to corona viride family which is present in the outer layer of corona virus. Long year ago, known corona viruses are HCOV-229E, NL63 etc., they have mild Symptoms such as general cold and cough. Less than 12-14 years third corona virus was discovered in the name of SARS-COV (Severe Acute Respiratory syndrome corona virus), MERS- COV (Middle East Respiratory Syndrome corona virus) and SARS-COV2. The WHO has announced that SARS- COV2 has launched a pandemic in early 2020. People infected with COVID-19 have a common fever, cold, cough while the pictures show the inflammation of lung. In some highly infected cases, it may cause severe damage to lung, multiple organ failure and death. The incubation period for coronavirus (COVID-19) is estimated to be between 1-14 days. The Government has proposed specific measures to protect people from the virus such as hand washing, social isolation, over- splitting.
2. Current Testing Methods RT-PCR TEST
In the current emergency, Reverse Transcription Polymerase Chain Reaction (RT-PCR) is the diagnostic method, used to identify the infection in humans. The test was performed using a nasal swab. A nasal swab, also called nasopharynx that detects viruses and bacteria that cause its respiratory infections. Inside the swab it contain droplets where there are viruses. All viruses have their own genetic material, here it is RNA (ribonucleic acid) and it is single stranded (ssRNA). In RT-PCR, PCR is a method to amplify double stranded DNA. Double stranded is nothing but which gives the multiple copies of DNA. Generally RNA is prepared from DNA that process is called transcription. Here DNA is prepared from RNA so this test process is called Reverse transcription. Once the process starts it produces multiple copies of DNA again and again. Then the fluorescent green started to produce in infected cells. RT-PCR test gives 98% accurate results and 2% false result. RT-PCR is not a diagnostic method to coronavirus. The below fig-1 and fig-2 shows the results of RT-PCR test based up on fluorescent green.
5488
Fig. 1. RT-PCR Test Positive
Fig. 2. RT-PCR Test Negative Blood Test
Blood test also called as rapid test. This type of test is similar to sugar test kit. We are going to use one drop of blood in the kit. The kit contains three levels C (control), IgG, IgM (both are antibodies) and also two places for buffer and blood. Buffer is nothing but a solution which rises up the blood up to control. Buffer solution is provided along with the test kit. When the process starts, after 10-15 minutes, a line visible in control place. There are two cases in this kit.
Case-1: if the line is visible in IgG: The person affected by the virus n number of times and he/she have
separate antibodies against each virus.
Case-2: if the line is visible in IgM: The person is affected by the virus in recent times which means before
one or two weeks and also it notifies that he/she must be take self quarantine with the help of doctors consultation.
Chest Findings
Chest CT, which shows a image of lungs in a clear way. It has also been used as a diagnostic tool. Chest CT can give clear a view of the region of the infected lung. Other biological diseases such as GGO, multifocal Pulmonary embolism and internal mutations and circulation. Using a CT scan the radiologist can easily distinguish whether the test result is positive or negative for COVID-19. Both tests are time consuming processes.
3. Existing System
Author proposed a process to obtain COVID-19 in early stage using data science and AI-Artificial Intelligent [1]. They utilized radiographic image(x- ray)and CAT image to investigate the soundness of a patients lungs. They suggested a way to take a Image extraction method to identify covid- 19 positive. Chest CT and x-ray images play an important role in this method. Initially datasets of images are collected and arranged. Collected images sizes are changed for faster algorithmic thinking. Then the greyscale images are converted to RGB images because reading of greyscale images in AI is little complex. Finally decision tree classifier method is used with Artificial intelligent to find whether the lung image is normal or covid-19. Deep learning technology is also used in this paper. At the initial stage datasets of images are collected and deep learning technology is
5489
also block diagram in fig 5.
(a) (b) (c)
Fig. 3. Chest X-ray COVID-19 Positive with AI(a) Actual Normal, Predicted Normal. (b) Actual COVID-19
+ve, Predicted COVID-19 Positive (c) Actual COVID-19 +ve, Predicted COVID-19 +ve
(a) (b) (c)
Fig. 4. Chest X-ray COVID-19 Positive with AI (a) Actual COVID-19+ve, Predicted COVID-19 +ve (b) Actual
COVID-19+ve, Predicted Normal (c) Actual COVID-19+ve, Predicted COVID-19 +ve
Fig. 5. Block Diagram to Find COVID-19 Using Data Science and AI
A system was proposed to inquire the diagnostic value and compatibility of chest CT. The dynamic modification of RT-PCR result was analyzed. serial chest CT scans are compared with basic test images. CT scan result is more sensitive than RT-PCR in detecting covid-19. RT-PCR procedure failed to detect the initial positive cases, whereas chest CT scan can easily detect. With the low collection of sample images, transportation and low kit performance, the total +ve rate of RT-PCR test was reported to be approximatively 30-60% at initial stages Author concluded that chest CT scan can detect covid-19 positive cases very sensitively than RT-PCR in short period. Fig 6 shows the flow diagram.
5490
Fig. 6. Study Flowchart COVID-19 = coronavirus 2019, RT-PCR =reverse-transcription Polymerase Chain
Reaction
Author in [3] approached a system to measure common and different manifestations of CAT image of three types of viruses using a systemic process. chest CAT scan features for all age group patients with COVID-19, SARS and MERS pneumonia are saved for further usage. Given datas Quality was evaluated in the way of literature and completeness. GGO, Lesion patterns, consolidation of three diseases of patients were recorded in the early stage. The below fig-7 shows the study flow.
105 patients
with –ve CT and +ve RT- PCR results 1049 patients suspected of COVID-19
underwent both chest CT and RT-PCR assays from January 6- February 6
Included patients
Excluded 35 patients: time interval of CT and RT-PCR
longer than 7 days
580 patients with +ve CT and +ve RT- PCR results 308 patients with +veCT and –ve RT- PCR results
28 patients with –ve CT and +ve RT- PCR
5491
Fig. 7. Flow Diagram of Study Selection
A system was proposed using deep features to identify coronavirus in the human body[4]. Author used covid19 (+ve) and (-ve) x-ray images as a reference. The covid19 (+ve) images are collected from Github repository and (-ve) images from kaggle. The deep CNN models are proposed to identify virus. Deep feature extraction methods are used. Author achieved 95.38% of accuracy in the result. From a particular layer, the deep features of Convolutional neural network models are extracted. Then the feature vector is obtained. The above figure shows the detection diagram fig-8.
A system was approached to identify SARS- COV2 infection from red blood cell counts using AI and machine language[5]. Authors collected a database about the infected people from some respective hospitals. Data contains people age
,
individual blood count, platelets etc., They approached random forecast and lasso-elastic-net regularized linear model is used to overcome lasso method. ANN (Artificial neural networks) model is evaluated. Also authors given the statistical analysis of blood counts. By using this method they predicted SARS-COV2 86% among infected people and 95% for regular ward patients. R Vaishya et.al and many others from new delhi, given an deep research article about the applications of AI in covid19[6]. Currently there is no drug or vaccine against virus. The only thing that helps people to avoid virus is to prevent themselves from coronavirus. On the other hand, finding coronavirus at the early stage will be useful for the treatment. RT-PCR and CT scan tests are using for the diagnosis of covid19, but it takes 1- 14days for the result.Studies Identified through database Searching
Number of Studies after duplicates removed
Studies screened by studies and abstract
Full text article assessed by eligibility
Studies included in the study
Studies excluded for theme irrelevance
Articles excluded for incomplete data or non applicable data retrieval and unmatched study
5492
Fig. 8. Detection of Corona Virus by SVM based on Deep Feature using X-ray Images
Artificial Intelligent, an upcoming technology in all fields. Artificial intelligent can analyze the CT scan readings in a better way. Not only for covid-19 AI can differentiate covid19 from some other respiratory illness. Early detection of infection is an important application of AI. Haleem A-n et al., presented the article about COVID-19 pandemic. Also the consequence of COVID-19 in day today life[7]. They are divided in to various categories like Healthcare, Social, and Economic. These three categories plays an important role in world trade. Also author discussed briefly about the categories mentioned above.
References
1. Dasari Naga Vinod, S.R.S. Prabaharan. Data Science and the role of Artificial Intelligence in achieving the fast diagnosis of Covid-19.
2. Tao Ai, MD, Zhenlu Yang, MD, Hongyan Hou MD, Chenao Zhan, MD, Chong Chen MD, Wenzhi LV BS, Qian Tao, Ziyong Sun, MD, Liming Xia, MD. Correlation of chest CT and RT-PCR testing for Coronavirus disease 2019 in china., Radiology 2020: 296: E40. https://doi.org/10.1148/radiol.2020200642
3. Xu Chen, Gang Zhang, Shuaiying Hao, Lin Bai, Jingjing Lu. Similarities and Differences of Early Pulmonary CT Features of Pneumonia Caused by SARS-COV-2, SARS-COV and MERS-COV: Comparison based on a Systemic review.
4. Prabira Kumar Sethy, Santi Kumara Behera. Detection of Corona virus Disease COVID-19 Based on Deep features.
5. Ahirup Banerjee, Surajit Ray, Bart Vorselaars, Joanne Kitson, Michail Mamalakis, Simonne Weeks, Mark Baker, Louise S. Mackenzie. Use of Machine Learning and Artificial Intelligence to Predict SARS-COV-2 infection from Full Blood Counts in a Population.
6. Raju Vaishya, Mohd Javaid, Ibrahim Haleem, Abid Haleem. Artificial Intelligence (AI) application forCOVID-19.
7. Haleem A-N et al., Effects of COVID-19 pandemic in daily life, current medicine research ad Practice. https://doi.org/10.1016/j:cmrp.2020.03.011
8. Ssairekha N, Kashwan KR, Improved Segmentation of MRI Brain Images by Denoising and Contrast Enhancement, Indian Journal of Science and Technology, Vol 8(22), DOI: 10.17485/ijst/2015/v8i22/73050, September 2015.
X ray image (Grey Scale Image)
3D Image (RGB)
Fully Connected Layer
Deep Features
SVM Classifier
