Classification of COVID-19 from Chest X-ray Images Using Transfer Learning

Lamyaa Alser Mohammed; Murtada Khalafallah Elbashir; Gais Alhadi Babikir

Lamyaa Alser Mohammed Department of Computer Sciences, University of Gezira, Wad Madani, Sudan
Murtada Khalafallah Elbashir Department of Computer Sciences, University of Gezira, Wad Madani, Sudan
Gais Alhadi Babikir Department of Computer Sciences, University of Gezira, Wad Madani, Sudan.

Abstract

Covid-19 descended from a virus strain called the corona, or coronavirus, in December 2019. It originated in Wuhan city of China, spread across the world, and became a pandemic. It should be noted that to prevent the further spread of this epidemic and to treat those infected with it quickly, researchers have been trying to develop effective methods to classify COVID2-19. One of these methods is the analysis of chest X-ray (CXR) images by radiologists, but this diagnosis has many drawbacks: it can take a long time, those radiologists are not available all the times, especially in remote areas, and the new features of COVID-19 are unknown to some radiologists. Therefore, it is necessary to implement an automatic classification system that helps doctors to classify COVID-19 in the early stages and provides a quick alternative diagnosis option to prevent the pandemic from spreading among people. This paper aimed to analyze and classify COVID-19-infected patients as infected (+ve) or not (−ve) from CXR images. It should be noted that the images used in this paper were collected from two sources as follows: 440 positive COVID-19 CXR images were collected from the GitHub repository, and 440 normal CXR images were obtained from the Kaggle repository. Then, data preprocessing techniques were applied, specifically resizing and normalization, to suit the classification process. Subsequently, three different transfer learning models (Xception, InceptionV3, and MobileNetV2) were proposed for the classification of coronavirus pneumonia-infected patients using chest CXR radiographs. Moreover, the experimental results obtained from the pre-trained Xception model have provided the highest classification performance with 99.43% accuracy, 99% precision, 99% recall, and 99% F1-score. It should be noted that, to the best of our knowledge, the best state-of-the-art model achieves an accuracy of 97%, precision of 99%, recall of 94 %, and F1-score of 91%. So, the results of the proposed model (Xception) are the best achieved so far. Therefore, the pre-trained Xception model could contribute great importance to the automatic classification of COVID-19 from CXR images.

References

[1] M. Ghaderzadeh and A. Farkhondeh, "Deep Learning in the Detection and Diagnosis of COVID-19 Using Radiology Modalities: A Systematic Review," Journal of Healthcare Engineering, p. 10, 2021.
[2] T. Ozturk, M. Talo, E. ,. Yildirim, U. ,. Baloglu, O. Yildirim and R. Acharya, "Automated detection of COVID-19 cases using deep neural networks with X-ray images," p. 11, 2020.
[3] A. K. Das, G. Sayantani, T. Samiruddin, D. Rohit, A. Sachin and C. Amlan, "Automatic COVID 19 detection from X ray images using ensemble learning with convolutional neural network," Pattern Analysis and Applications, p. 14, 2021.
[4] P. Saha, M. ,. Sadi and M. M. Islam, "EMCNet: Automated COVID-19 diagnosis from X-ray images using convolutional neural network and ensemble of machine learning classifiers," Informatics in Medicine Unlocked, p. 12, 2021.
[5] I. Baltruschat, H. Nickisch, M. Grass, T. Knopp and A. Saalbach, "Comparison of Deep Learning Approaches for Multi-Label Chest X-Ray Classifcation," nature, 2019.
[6] T. Krishna and K. ,. Hemantha, "Deep Learning and Transfer Learning Approaches for Image Classification," International Journal of Recent Technology and Engineering (IJRTE), p. 7, 2019.
[7] E. Goldstein, D. Keidar, D. Yaron, Y. Shachar, A. Blass, L. Charbinsky and I. Aharony, "COVID-19 Classification of X-ray Images Using Deep Neural Networks," p. 22, 2020.
[8] S. Asif, Y. Wenhui, H. Jin and S. Jinhai, "Classification of COVID-19 from Chest X-ray images using Deep Convolutional Neural Networks," p. 7, 2020.
[9] E. E.-D. Hemdan, M. A. Shouman and M. E. Karar, "COVIDX-Net: A Framework of Deep Learning Classifiers to Diagnose COVID-19 in X-Ray Images," p. 14, 2020.
[10] A. Narin, K. Ceren and P. Ziynet, "Automatic Detection of Coronavirus Disease (COVID-19) Using X-ray Images and Deep Convolutional Neural Networks," p. 17, 2020.
[11] M. Rahimzadeh and A. Abolfazl, "A modified deep convolutional neural network for detecting COVID-19 and pneumonia from chest X-ray images based on the concatenation of Xceptionand ResNet50V2," Informatics in Medicine Unlocked, p. 9, 2020.
[12] J. P. Cohen, P. Morrison and L. Dao, "COVID-19 image data collection," 2020. [Online]. Available: http://arxiv.org/abs/2006.11988.. [Accessed 12 4 2021].
[13] P. Mooney, "Chest X-ray images (pneumonia)," Kaggle Repository, 2018. [Online]. Available: https://www.kaggle.com/paultimothymooney/chest-xray-pneumonia, 2018. [Accessed 12 4 2021].
[14] A. M. Ismael and A. Şengür. Deep learning approaches for COVID-19 detection based on chest X-ray images. Expert Systems with Applications, 164, 114054, 2021.
[15] A. Abbas, M. M. Abdelsamea and M. M. Gaber. Classification of COVID-19 in chest X-ray images using DeTraC deep convolutional neural network. Applied Intelligence, 51(2), 854-864, 2021.
[16] R. Abdalrhman, M. Elbashir and G. A.Babikir, "Classification of COVID-19 in Chest X-Ray Images using Deep Transfer Learning," Gezira Journal of Engineering and Applied Sciences, 16(1), 1-6, 2022.