COVID-19 Detection in Chest Radiograph Based on YOLO v5

ABSTRACT

The current pandemic of COVID-19 has brought certain difficulties to the detection and diagnosis. With the continuous development of medical imaging technology, chest radiograph has become a common examination method for detecting lung diseases. Reasonable use of new coronary pneumonia chest radiographs and machine learning related algorithms to achieve efficient, accurate and automatic identification of covid-19 is extremely important. Based on the detection of four types of COVID-19, this paper proposes a method for the detection and classification of COVID-19 based on the YOLOv5 model. Experimental results show that our algorithm has the best performance compared with other deep learning algorithms. Specifically, the map@0.5 index of the prediction result of our algorithm model is 0.605, which is 32.096% and 18.627% larger than the Fast RCNN algorithm and the Efficient Net model respectively. © 2021 IEEE.