Detection of COVID-19 coronavirus infection in chest X-ray images with deep learning methods

Early detection of patients with COVID-19 coronavirus infection is essential in ensuring an adequate treatment and reducing the burden on the health care system. An effective method of detecting COVID-19 is computer analysis of chest X-rays. The paper proposes a methodology that consists of stages of formatting X-ray images to the size (224, 224) size, their classification using deep convolutional neural networks, such as Xception, InceptionResnetV2, MobileNetV2, Dense-Net121, ResNet50 and VGG16, which are pre-trained on the ImageNet dataset and then fine-tuned on a set of chest X-rays. The results of computer experiments showed that the VGG16 model with fine-tuning of parameters demonstrated the best performance in the COVID-19 classification with accuracy = 99.09%, recal = 99.483%, precision = 99.08% and f1_score = 99.281%.