A Computer-Assisted Diagnosis Tool for Classifying COVID-19 based on Chest X-Ray Images

Objective: Since COVID-19 is a worldwide pandemic, detection using convolutional neural network (CNN) has been an extraordinary research technique. In the reported studies, many models that can predict based on deep learning methods various medical images have created; however, clinical decision support systems limited. The aim of this study to develop successful model X-ray and computer-assisted, fast, free web-based diagnostic tool for accurate COVID-19.
 Method: 15-layer CNN was used detect images, which outperformed previously published in terms classification. performance evaluated according Accuracy, Matthews Correlation Coefficient (MCC), F1 Score, Specificity, Sensitivity (Recall), Youden’s Index, Precision (Positive Predictive Value: PPV), Negative Value (NPV), Confusion Matrix (Classification matrix). second phase study, computer-aided disease developed Python Flask library, JavaScript Html codes.
 Results: diagnose average accuracy 98.68 % training set 96.98 testing set. Among evaluation metrics, minimum value 93.4 MCC index, maximum 97.8 sensitivity NPV. A higher means lower false negative (FN) value, low FN encouraging outcome cases. This conclusion crucial because minimizing overlooked cases (false negatives) one main goals research. 
 Conclusion: period when spreading rapidly around world, it thought X-Ray be very effective fast tool. system assist physicians radiologists making decisions about disease, as well provide diagnosis, follow-up, prognosis. computer-assisted diagnosis publicly accessed at http://biostatapps.inonu.edu.tr/CSYX/.