Fast and Accurate Detection of Covid-19-related Pneumonia from Chest X-ray Images with Novel Deep Learning Model

Background: Novel coronavirus disease has spread rapidly worldwide. As recentradiological literatures on Covid-19 related pneumonia is primarily focused onCT findings, the American College of Radiology (ACR) recommends using portablechest X-radiograph (CXR). A tool to assist for detection and monitoring ofCovid-19 cases from CXR is highly required. Purpose: To develop a fullyautomatic framework to detect Covid-19 related pneumonia using CXR images andevaluate its performance. Materials and Methods: In this study, a novel deeplearning model, named CovIDNet (Covid-19 Indonesia Neural-Network), wasdeveloped to extract visual features from chest x-ray images for the detectionof Covid-19 related pneumonia. The model was trained and validated by chestx-rays datasets collected from several open source provided by GitHub andKaggle. Results and Discussion: In the validation stage using open-source data,the accuracy to recognize Covid-19 and others classes reaches 98.44100model to classify Covid-19 and other pathologies might slightly decrease theaccuracy. Although SoftMax was used to handle classification bias, thisindicates the benefit of additional training upon the introduction of new setof data. Conclusion: The model has been tested and get 98.4source datasets, the sensitivity and specificity are 100respectively.