Research Article Open Access

Classification of Thoracic X-Ray Images of COVID-19 Patients Using the Convolutional Neutral Network (CNN) Method

Ramacos Fardela1, Dian Milvita1, Mawanda Almuhayar2, Dedi Mardiansyah1, Latifah Aulia Rasyada1 and Lukman Mul Hakim3
  • 1 Department of Physics, Faculty of Mathematics and Natural Sciences, Universitas Andalas, Indonesia
  • 2 Department of Mathematics and Data Science, Faculty of Mathematics and Natural Sciences, Universitas Andalas, Indonesia
  • 3 Department of Computer Engineering, Faculty of Information Technology, Universitas Andalas, Indonesia

Abstract

Recently, radiology modalities have been widely used to detect COVID-19. Thoracic X-rays and CT scans are the primary radiological tools utilized in the diagnosis and treatment of individuals with COVID-19. In addition, chest CT scans are more accurate and sensitive in early COVID-19 identification. A new problem arises in diagnosing the results of CT scan images of COVID-19 by radiologists or radiology specialists where COVID-19 is difficult to distinguish from pneumonia caused by other viruses and bacteria, so misdiagnosis can occur. Many researchers worldwide have developed computer-aided detection or diagnosis schemes based on medical image processing and machine learning to overcome this challenge. This research focuses on the development of previous studies, where the use of the Convolutional Neural Network (CNN) method to classify Thoracic X-ray Images of COVID-19 Patients is compared with the model developed by Roboflow. Image manipulation techniques applied to this study are pseudo color and the program is Python. This study employs the pseudo color image manipulation technique of the program in Python. This study uses data on patients with confirmed COVID-19 at Andalas University Hospital in 2022. Based on the study's results, a very good CNN Specificity score of 93% was obtained and the perfect Sensitivity score value was produced by the detection method using the Roboflow model, which was 100%. However, the Kappa score for both methods is below the expected threshold of 36-38%. Based on the ROC value, the CNN and Roboflow methods are good for calculating chest X-ray images of COVID-19 and normal patients.

Journal of Computer Science
Volume 20 No. 4, 2024, 357-364

DOI: https://doi.org/10.3844/jcssp.2024.357.364

Submitted On: 17 November 2023 Published On: 1 February 2024

How to Cite: Fardela, R., Milvita, D., Almuhayar, M., Mardiansyah, D., Rasyada, L. A. & Hakim, L. M. (2024). Classification of Thoracic X-Ray Images of COVID-19 Patients Using the Convolutional Neutral Network (CNN) Method. Journal of Computer Science, 20(4), 357-364. https://doi.org/10.3844/jcssp.2024.357.364

  • 814 Views
  • 399 Downloads
  • 0 Citations

Download

Keywords

  • Convolutional Neural Network
  • COVID-19
  • Early Detection
  • Roboflow