Deep Learning-Based Convolutional Neural Network for COVID-19 and Pneumonia Prediction

Severe Acute Respiratory Syndrome and COVID-19 is a harmful illness which can be transmitted by small particles of saliva or mucus that evaporate whenever a sick person sneezes or coughs. It spreads quickly through close interaction with a person who has been infected, or contact with surfaces or objects that have been already contaminated with the virus. Pneumonia is another infectious disease which can be brought on by an infection from a bacterium in the lungs' alveoli. The occurrence of inflamed infected lung tissue shows the production of pus. Experts execute physical assessments on their patients and use various tests such as chest X-rays, ultrasounds, and lung biopsies to identify whether the patient has any disease. The late diagnosis and misdiagnosis of the disease may lead to death. Nowadays, the development of deep learning is assisting doctors in diagnosing patients with COVID-19 or pneumonia and many other lifethreatening diseases. The proposed work uses X-ray images substituted to a CNN-based deep learning approach for predicting whether a patient is affected by the covid-19 or pneumonia. The proposed model has been developed by using a dataset containing 9,208 images. During the process of model development, the results have been demonstrated in accuracy as 98%. This work is capable of determining and predicting four different classes namely COVID-19, bacterial pneumonia, viral pneumonia, and normal lungs in the chest X-ray images dataset.