Covid-19 infected lung detection using machine learning

Abstract

In every 100 years, there has been a pandemic all around the world. The globe faced Plague, Cholera, and Spanish Flu in the years 1720, 1820, and 1920, respectively. Coronavirus, commonly known as Covid-19, is currently circulating in 2020. Coronavirus affects the nose, sinuses, upper neck, and lungs, among other parts of the human respiratory system. Coronaviruses come in a variety of types, although the majority of them aren’t harmful. A brand-new coronavirus epidemic occurred in the Chinese city of Wuhan in December 2019. It was first recognized as SARS-CoV-2 by the World Health Organization, then renamed Covid-19, and it spread swiftly over the world by March 2020. The novel COVID-19 has the potential to develop an infection of the respiratory system. In both the upper and lower respiratory tracts, it can affect the sinuses, nose, throat, windpipe, and lungs.COVID19 is a virus that infects humans via respiratory droplets, coming into contact with a positive for COVID19patientCOVID-19 detection is one of the most challenging undertakings in the globe owing to the virus’s fast spread. The number of persons diagnosed with COVID-19 is increasing dramatically, according to data, with over 16 million confirmed cases. For our research, we’re looking for COVID-19 symptoms in patients’ chest X-ray pictures. We began by gathering information from a variety of sources and categorizing it as COVID-19 positive, other lung illnesses, and normal chest X-ray pictures. Second, we used VGG16, InceptionV3, and ResNet50 to classify the data. The accuracy rates for VGG16, ResNet50, and InceptionV3 were 97.82 percent, 98.89 percent, and 97.65 percent, respectively. Then we combined these classifiers into an ensemble model, and COVDet19 V1 attained an overall accuracy of 97.92 percent.