COVID-19 CASES DETECTION FROM X-RAY IMAGES BY USING DEEP NEURAL NETWORKS
DOI:
https://doi.org/10.24867/16OI04OrcicKeywords:
COVID-19 classification, convolutional neural networks, deep learningAbstract
The COVID-19 pandemic has a destructive effect on health on the population worldwide. The crucial step in the fight against it is to control the spread of the disease while screening the broad number of suspected cases for appropriate quarantine and treatment as priority. Capacity and quality of laboratory testing is a challenging task so alternative methods in testing plays a significant role in this fight. Based on that, one of the novel approaches relies on analyzing radiological imaging using chest radiography for diagnosis, assessment and staging of COVID-19 infection. Developing an automated tool that will utilize the large number of X-rays for classification would be of great importance when covering a large number of cases. In the previous years, the state-of-the-art convolutional neural networks architectures showed outstanding results in numerous medical classification tasks. Motivated by this, experiments conducted in this paper are analyzing their usage in the task of detection of COVID-19 cases by X-ray image classification along with employing the transfer learning strategy on pre-trained grayscale ImageNet while also using various pre-processing techniques. The results presented in this paper come to the conclusion that deep convolutional neural networks could extract radiological visual features that correlates with biomarkers that are related to COVID-19 cases with the high accuracy.
References
[2] Mehta, P., McAuley, D. F., Brown, M., Sanchez, E., Tattersall, R. S., Manson, J. J., & HLH Across Speciality Collaboration. (2020). COVID-19: consider cytokine storm syndromes and immunosuppression. Lancet, 395(10229), 1033.
[3] Chung, M., Bernheim, A., Mei, X., Zhang, N., Huang, M., Zeng, X., ... & Jacobi, A. (2020). CT imaging features of 2019 novel coronavirus (2019- nCoV). Radiology, 295(1), 202-207.
[4] Bar, Y., Diamant, I., Wolf, L., & Greenspan, H. (2015, March). Deep learning with non-medical training used for chest pathology identification. In Medical Imaging 2015: Computer-Aided Diagnosis (Vol. 9414, p. 94140V). International Society for Optics and Photonics.
[5] Bar, Y., Diamant, I., Wolf, L., Lieberman, S., Konen, E., & Greenspan, H. (2015, April). Chest pathology detection using deep learning with non-medical training. In 2015 IEEE 12th international symposium on biomedical imaging (ISBI) (pp. 294-297). IEEE.
[6] Greenspan, H., Van Ginneken, B., & Summers, R. M. (2016). Guest editorial deep learning in medical imaging: Overview and future promise of an exciting new technique. IEEE Transactions on Medical Imaging, 35(5), 1153-1159.
[7] Tartaglione, E., Barbano, C. A., Berzovini, C., Calandri, M., & Grangetto, M. (2020). Unveiling COVID-19 from Chest X-ray with deep learning: a hurdles race with small data. arXiv preprint arXiv:2004.05405.
[8] ChestXRay2017, https://data.mendeley.com/datasets/rscbjbr9sj/2/files/f12eaf6d6023-432f-acc9-80c9d7393433 (pristupljeno u julu 2020.)
[9] Bloice, M. D., Roth, P. M., & Holzinger, A. (2019). Biomedical image augmentation using Augmentor. Bioinformatics, 35(21), 4522-4524.
[10] He, K., Zhang, X., Ren, S. and Sun, J., 2016. Deep Residual Learning for Image Recognition.
[11] COVID-Net Open Source Initiative, https://github.com/lindawangg/COVID-Net (pristupljeno u julu 2020.)
