The coronavirus, which appeared in Wuhan city of China and named COVID-19 , spread rapidly and caused the death of many people.Early diagnosis is very important to prevent or slow the spread.The first preferred method by clinicians is real-time reverse tommy todd ointment transcription-polymerase chain reaction (RT-PCR).However, expected accuracy values cannot be obtained in the diagnosis of patients in the incubation period.Therefore, common lung devastation in COVID-19 patients were considered and radiological lung images were used to diagnose.
In this study, automatic COVID-19 diagnosis was made from posteroanterior (PA) chest X-Ray images by deep learning method.In the study, using two different deep learning methods, classification was made with different dataset combinations consisting of healthy, COVID, bacterial pneumonia and viral pneumonia X-ray images.The results show that the proposed deep learning-based donitsirauta system can be used in the clinical setting as a supplement to RT-PCR test for early diagnosis.