COVID-19 detection with X-ray images by using transfer learning

Coronavirus is an infectious disease induced by extreme acute respiratory syndrome coronavirus 2. Novel coronaviruses can lead to mild to serious symptoms, like tiredness, nausea, fever, dry cough and breathlessness. Coronavirus symptoms are close to influenza, pneumonia and common cold. So Coronavirus can only be confirmed with a diagnostic test. 218 countries and territories worldwide have reported a total of 59.6 million active cases of the COVID-19 and 1.4 million deaths as of November 24, 2020. Rapid, accurate and early medical diagnosis of the disease is vital at this stage. Researchers analyzed the CT and X-ray findings from a large number of patients with coronavirus pneumonia to draw their conclusions. In this paper, we applied Support Vector Machine (SVM) classifier. After that we moved on to deep transfer learning models such as VGG16 and Xception which are implemented using Keras and Tensor flow to detect positive coronavirus patient using X-ray images. VGG16 and Xception show better performances as compared to SVM. In our work, Xception gained an accuracy of 97.46% with 98% f-score. © 2022 - IOS Press. All rights reserved.

COVID-19 Patients Prediction Based on Symptoms Using Fuzzy Logic Approach

Coronavirus is a contagious disease that has frightened the globe and continued to threaten the livelihoods of millions of individuals. The detection of COVID-19 has lately become a critical task for medical professionals. Researchers and experts from many areas are working tirelessly to discover preventive ways to preserve the earth from this unseen virus. In disease studies, artificial intelligence approaches have proven to be successful. A Type-2 Fuzzy Logic method has been designed in this work to help in the preliminary diagnosis of whether a patient's symptoms are likely related to a COVID-19 infection. Statistics and information on the rule-based method were obtained from publicly available datasets and databases. Based on the symptoms that the patient exhibits, the algorithm infers the probability of coronavirus contamination. This proposed automated inference method can aid doctors in recognizing diseases and individuals in self-diagnosing. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.