Article Preview
Top1. Introduction
The coronavirus has been confirmed as Wuhan, Hubei's biggest metropolis, was the site of the first official case in China, according to WHO (Ai et al.,2020). Thousands of individuals have been perished, and there have been millions of confirmed cases across the world. As a result, it took the form of a pandemic, wreaking havoc on the health and well-being of humanity's entire population. This has resulted in the SARS (severe acute respiratory syndrome) which is a condition in which the (SARS-CoV) virus causes coronavirus illness, which is known as coronavirus disease COVID-19 (acronym). The coronavirus, like SARS and MERS, is a member of the coronavirus family (Shelke et al.,2021) . This dangerous sickness travels around the world considerably faster than regular well-known flu. This research tries to illustrate that viruses in this family have severe symptoms. This research observed that viruses in this family are prominently displayed in radiography photographs, and it gets more complex to identify virus’s evolution. As a result, classification of Covid-19 disease using chest X-rays can be determined (CXR) (Ahmed et al.,2020). In order to speed up the process of research and development and create new norms and standards to help treat those affected by the coronavirus pandemic, WHO is bringing together scientists from around the world and experts in global health. The CDC advises using the majority of diseases that are susceptible to pandemics, including coronavirus, may be accurately detected using polymerase chain reaction (PCR), the gold standard in molecular diagnostics (Singh et al.,2021).
The only known issue with COVID-19 detection Techniques is clinical testing kits and the experts may not always be available, especially in remote places. So, using chest X-ray CXR pictures, we proposed a simple and low-cost machine learning-based technique for classifying COVID-19 +ve and –ve cases. This method can detect COVID-19 positive patients with near-perfect precision. As part of this study, we provided a strategy for detecting COVID-19 positive or negative patients (Babukarthik et al.,2020) We've demonstrated that the proposed tool is effective in terms of classification accuracy and sensitivity. This study uses data from open sources from several data-based websites (Jamshidi et al.,2020) . When combined with some unsupervised approaches to machine learning, the suggested COVID-19 detection method utilizing machine learning can increase diagnosis accuracy and decrease diagnosis error. The CNN and GNN (Narin et al., 2021) models have been updated for the discovery of COVID-19(Huang et al.,2020) cases from chest X-ray by adding a clustering method to the dataset. Images from chest X-rays can also be utilized to diagnose cancer.
Artificial intelligence is one of the most beneficial technologies for diagnosing COVID-19 infections from chest X-rays is machine learning. In previous research-works done by several authors X-rays of the chest helped us a lot. It was utilized to identify COVID-19, with 71 COVID-19 and 5000 non-COVID-19 pictures (Aggarwal et al.,2020). The results were based on a database of 5071 chest X-rays (Aggarwal et al.,2020). The utilization of chest X-ray pictures over CT of the thorax utilizing multiple machine learning techniques to boost COVID-19 detection of the patients has been proposed in this study. As a result, as compared to existing models, our suggested deep learning algorithms provide improved COVID-19 identification accuracy and require less time to find patients. By using the concept of Transfer Learning the accuracy can be improved significantly. We have utilized a VGG16 convolutional neural network which is pre-trained using a portion of the 14 million+ picture ImageNet dataset, which is organized into 22,000+ categories. In 2015 (Simonyan et al.,2020) suggested this concept. Some researchers preferred to use Deep CNN-LSTM architecture – a hybrid model to classify images which is implemented in our work.