Detection of COVID-19 from Chest X-Ray Images Using Deep Neural Network with Fine-Tuning Approach

The coronavirus (COVID-2019) quickly spread throughout the world and came to be a pandemic. To avoid further spreading this epidemic and treat affected patients rapidly, it is important to recognize the positive cases as early as possible. In this paper, deep learning techniques are employed to detect COVID-19 from chest X-ray images quickly. The images of the two classes, COVID and No-findings are collected from three public datasets. The proposed approach consists of two phases; transfer learning and fine-tuning. Transfer learning is carried out by seven deep learning models: DenseNet, Inception_Resnet, MobileNet, NASNet, ResNet, VGG, and Xception. Two fully-connected layers are added to the pre-trained model for fine-tuning. These models' performance is compared in terms of accuracy, sensitivity, specificity, and computation time. The experimental results showed that MobileNet obtained 98%, which outperformed all other accuracy and time models. © 2021, The Author(s), under exclusive license to Springer Nature Switzerland AG.