Implementing a Hierarchically Classification of COVID-۱۹ and Pneumonia Diseases Using Deep Learning

In December ۲۰۱۹, the COVID-۱۹ virus, which was firstly observed in Wuhan, China, quicklyspread around the world and became a global pandemic. At the beginning of the disease outbreak,due to its unknown nature and similarity of the symptom with the general structure of the diseasesof the COVID group such as SARS and MERS, posed significant diagnostic challenges for physicians.So that the main diagnostic approach for physicians, considering the degree of lung involvement,was to observe the patients’ X-Rays images. By increasing of COVID-۱۹ disease andit’s becoming a major issue all over the world, a huge data with different topics and formats hasbeen produced in relation to the causes and different aspects of this disease, and as a result, it helpsthe growth of researches related to the disease. Meanwhile, artificial intelligence and machinelearning algorithms have played a very prominent role in diagnosis of this disease. In this paper,we present a deep convolutional neural network for diagnosis of COVID-۱۹ and pneumonia diseasesby using X-Rays images. For this purpose, a hierarchical classification is adopted by usingInceptionV۳ fully connected deep convolutional architecture and by training two similar architectureson COVID-۱۹, pneumonia and normal images for diagnosis of the diseases. The resultsof the proposed algorithm show an average accuracy of ۹۸.۷۶ % for classification of two diseasesCOVID-۱۹ and pneumonia which is a high accuracy in comparison with similar methods.

کلیدواژه ها:

COVID-۱۹ ، Pneumonia ، X-Rays Images ، Deep Neural Network ، InceptionV۳net ، Hierarchically classification