MCNN: A Deep Learning Based Rapid Diagnosis Method for COVID-19 from the X-ray Images

Due to the rapidly spreading nature of coronavirus, a pandemic situation has emerged around the world It is affecting society at large that includes the global economy and public health too It was found in recent studies that the novel and unknown nature of this virus makes it more difficult to identify and treat the affected patient in the early stage In this context, a time-consuming method named reverse transcription-polymerase chain reaction (RT-PCR) is being used to detect the positive cases of COVID-19, which requires blood samples of the suspects to diagnose the disease This paper presents a new deep learning-based method to detect COVID-19 cases using chest X-ray images as the recent studies show that the radiology images have relevant features that can be used to predict the COVID-19 The proposed method is developed for binary classification to identify that a person is infected with COVID-19 or not A total of 2400 X-ray images are taken for the experimental work It includes 1000, COVID-19, and 1000, non-COVID-19 images, 200, COVID-19, and 200, non-COVID-19 testing images The proposed method has been compared with the existing state-of-the-art methods on various statistical parameters which give better results with higher accuracy in diagnosing the COVID-19 cases The proposed method has obtained 98 25% accuracy, 98 49% precision, 98% sensitivity, 98 50% specificity, and 98 25% F1 score © 2020 Lavoisier All rights reserved