Performance Comparison Between Convolutional Neural Network-Based Artificial Intelligence SiCoSa with a Combination of 2 Radiology Specialists at Universitas Airlangga Hospital in Diagnosing COVID-19 on Chest X-Ray Images
Caesario Indra Ardana1, Anggraini Dwi Sensusiati2* and Niko Azhari Hidayat3
Abstract
Introduction: COVID-19 can be detected in various ways, namely RT-PCR as the gold standard which is relatively expensive and long or with imaging such as X-Ray which is relatively cheaper and fast. However, X-Ray for diagnosing COVID-19 has a drawback, namely for radiologists to interpret it. Artificial intelligence is here to address this issue of radiologist availability. Method: This study is an analytic observational study using a cross-sectional design based on secondary data. Sampling was done by non-probability sampling technique. The area of the AUROC curve used to determine a diagnostic tool. The wider the AUROC curve the better a diagnostic tool. Research Result: In this study, a sample of 500 images was obtained with 460 chest X-ray images true positive COVID-19 and 40 chest X-ray images true negative COVID-19. SiCoSa with MobileNetV2 architecture has sensitivity, specificity, and AUROC respectively by 100%, 0%, and 0.5 while SiCoSa with ResNet 50 architecture has a sensitivity, specificity, and AUROC respectively at 82.61%, 20%, and 0.5131. Interpretation both 2 Radiology Specialists at Universitas Airlangga Hospital have sensitivity, specificity, and AUROC respectively 69.35%, 100%, and 0.8467. Conclusion: SiCoSa has lower performance than the Interpretation of 2 Radiology Specialists at Universitas Airlangga Hospital in differentiating COVID-19 pneumonia and non-COVID-19 pneumonia.
Keywords
artificial intelligence; SiCoSa; COVID-19; chest x-ray image
Cite This Article
Ardana, C. I., Sensusiati, A. D., Hidayat, N. A. (2022). Performance Comparison Between Convolutional Neural Network-Based Artificial Intelligence SiCoSa with a Combination of 2 Radiology Specialists at Universitas Airlangga Hospital in Diagnosing COVID-19 on Chest X-Ray Images. International Journal of Scientific Advances (IJSCIA), Volume 3| Issue 6: Nov-Dec 2022, Pages 911-914, URL: https://www.ijscia.com/wp-content/uploads/2022/12/Volume3-Issue6-Nov-Dec-No.376-911-914.pdf
Volume 3 | Issue 6: Nov-Dec 2022
