Abstract
COVID-19 has ended up being the greatest pandemic to come to pass for on
humanity in the last century. It has influenced all parts of present day
life. The best way to confine its spread is the early and exact finding
of infected patients. Clinical imaging strategies like Chest X-ray
imaging helps specialists to assess the degree of spread of infection.
In any case, the way that COVID-19 side effects imitate those of
conventional Pneumonia brings few issues utilizing of Chest Xrays for
its prediction accurately. In this investigation, we attempt to assemble
4 ways to deal with characterize between COVID-19 Pneumonia,
NON-COVID-19 Pneumonia, and an Healthy- Normal Chest X-Ray images.
Considering the low accessibility of genuine named Chest X-Ray images,
we incorporated combinations of pre-trained models and data augmentation
methods to improve the quality of predictions. Our best model has
achieved an accuracy of 99.5216%. More importantly, the hybrid did not
predict a False Negative Normal (i.e. infected case predicted as normal)
making it the most attractive feature of the study.