A Comparative Analysis for the Detection of COVID-19 from Chest X-ray Dataset

Abstract
Corona Virus Disease (COVID-19) is probably one of the most dreadful disease globally and the root cause of people’s demise. Advance forms of COVID-19 are being discovered every time, and it has become hard to find solutions to get rid of this severe trouble. Scientists, Researchers, and others who work in medical fields are trying to implement a system that will find symptoms of COVID-19 within a short time so that necessary treatment can be taken before the disease becomes acute. Deep learning methods in that scenario are giving some promising results in detecting COVID-19 efficiently. Deep CNN (Convolutional Neural Network) architectures like ResNet, VGG-16, EfficientNet and other models are accomplishing excellent results in identifying COVID-19. In this paper, some works related to deep learning and COVID-19 by several researchers are discussed. Furthermore, four models were proposed, which are four versions of EfficientNet, and the models were trained on the largest dataset named COVIDx CXR-3, containing approximately 30,882 images, where 16690 are COVID positive and 14190 are COVID negative. Among the models, EfficientNet B3 provided a good enough accuracy of 97.89 % with a sensitivity of 98.02 % and a specificity of 97.67 %. In the end, EfficientNet B3 was compared with other CNN algorithms published by different researchers and found that the proposed model provided a more acceptable accuracy than the other models compared. Moreover, the methodology of the proposed models and their descriptions have been given in this paper.

BibTeX

@inproceedings{anik2022comparative,
  title={A Comparative Analysis for the Detection of COVID-19 from Chest X-ray Dataset},
  author={Anik, Anirban Saha and Chakraborty, Kowshik and Datta, Bishowjit and Kader, Abdul and Morol, MD Kishor},
  booktitle={2022 International Conference on Recent Progresses in Science, Engineering and Technology (ICRPSET)},
  pages={1--6},
  year={2022},
  organization={IEEE}
}

Recommended citation: Anirban Saha Anik, Kowshik Chakraborty, Bishowjit Datta, Abdul Kader, MD. Kishor Morol. "A Comparative Analysis for the Detection of COVID-19 from Chest X-ray Dataset." International Conference on Recent Progresses in Science, Engineering and Technology (ICRPSET 2022). DOI: 10.1109/ICRPSET57982.2022.10188570.
Download Paper