A new method of diagnosing Covid-۱۹ based on deep learning

The sudden outbreak of COVID-۱۹ has caused serious global concern due to the high rate of infections and deaths. Considering the limitation of test kits and the need for rapid screening and diagnosis of patients, it is necessary to perform an automatic diagnosis model as a rapid diagnosis system to detect the infection of COVID-۱۹ and prevent its spread among people, therefore, the use of convolutional neural network algorithms (CNN) based on deep learning to screen chest X-ray images (CXR) of patients with covid-۱۹, have become more necessary and have taken a major share of articles in this field. Paying attention to the highly infectious nature of this virus, correct, accurate and timely diagnosis of Covid-۱۹ disease in people infected with this virus is very necessary to curb the global spread of this disease. Despite the remarkable progress in the field of diagnosis of this disease, in some cases the results are controversial, so this research examines a new method in diagnosis.In this article, deep learning architectures are used to diagnose this disease. The combination of two Depthwise Separable Convolution algorithms with VGG۱۹, which had better performance with ۹۹% accuracy.

Keywords:

Covid-۱۹ , convolution network , DepthWise Separable Convolution , VGG۱۹.