Identification of Pneumonia in Corona Virus Patient

A novel corona virus is dangerous and life threatening as patient's lungs are affected very severely by pneumonia. When found in x-ray the chest was fully dense and made lungs rubber type. This pneumonia was quite different from normal pneumonia. In Corona virus 19 people's lungs were so much affected that they were unable to breathe and their oxygen level declined very rapidly and the nebulizer didn’t work for them as it is more effective in normal pneumonia patients. In this paper we have taken 2000 dataset to study the difference between person suffering from lung disease chest x-ray/ corona virus patient chest x-ray and pneumonia patient chest x-ray by using deep learning. With the help of chest X-ray (CXR) pictures, system categorizes corona virus into pneumonia, non- pneumonia, and healthy images. 1468 CXR pictures were acquired from publicly available datasets, consisting of 222, 683, and 700 CXR images of corona virus pneumonia, non-pneumonia, and healthy samples, respectively. The system used VGG16 model and a mixture of traditional and approaches for data augmentation. The VGG16-based model was compared to other types of models. The Corona Automatic diagnosis x-ray images (CADx) system was built and evaluated. A test set of 192 CXR pictures was used to assess three-category accuracy. This system obtained 83.6 percent of accuracy over chest x-rays to classify corona virus pneumonia, non-pneumonia and healthy patient. The sensitivity was above 95%.